cavis/libnd4j/include/helpers/Loops.h

1494 lines
68 KiB
C
Raw Normal View History

2019-06-06 14:21:15 +02:00
/*******************************************************************************
* Copyright (c) 2015-2018 Skymind, Inc.
*
* This program and the accompanying materials are made available under the
* terms of the Apache License, Version 2.0 which is available at
* https://www.apache.org/licenses/LICENSE-2.0.
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
* License for the specific language governing permissions and limitations
* under the License.
*
* SPDX-License-Identifier: Apache-2.0
******************************************************************************/
//
// @author Yurii Shyrma (iuriish@yahoo.com), created on 14.03.2019
//
2019-06-06 14:21:15 +02:00
#ifndef LIBND4J_LOOPS_H
#define LIBND4J_LOOPS_H
#include <functional>
#include <system/pointercast.h>
#include <helpers/shape.h>
#include <helpers/LoopKind.h>
#include <helpers/OmpLaunchHelper.h>
#include <array/DataTypeUtils.h>
#include <ops/ops.h>
#include <loops/indexreduce.h>
2019-06-06 14:21:15 +02:00
#include <helpers/ConstantTadHelper.h>
#include <system/openmp_pragmas.h>
#include <execution/Threads.h>
2019-06-06 14:21:15 +02:00
namespace sd {
2019-06-06 14:21:15 +02:00
template <typename X, typename Z, typename E>
class ND4J_EXPORT ReductionLoops {
protected:
public:
template <typename OpType>
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
static FORCEINLINE void loopReduce(sd::memory::Workspace* workspace, const X* x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int* dims, E* extraParams);
2019-06-06 14:21:15 +02:00
};
template <typename X, typename Z>
class ReductionFloatLoops : public ReductionLoops<X, Z, Z> {
2019-06-06 14:21:15 +02:00
public:
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
static void wrapper(int opNum, sd::memory::Workspace* workspace, const X* x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int* dims, Z* extraParams);
2019-06-06 14:21:15 +02:00
template <typename OpType>
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
static void innerloopReduce(sd::memory::Workspace* workspace, const X* x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int* dims, Z* extraParams);
2019-06-06 14:21:15 +02:00
};
template <typename X, typename Z>
class ND4J_EXPORT ReductionBoolLoops : public ReductionLoops<X, Z, X> {
2019-06-06 14:21:15 +02:00
public:
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
static void wrapper(int opNum, sd::memory::Workspace* workspace, const X* x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int* dims, X* extraParams);
2019-06-06 14:21:15 +02:00
template <typename OpType>
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
static void innerloopReduce(sd::memory::Workspace* workspace, const X* x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int* dims, X* extraParams);
2019-06-06 14:21:15 +02:00
};
template <typename X, typename Z>
class ND4J_EXPORT ReductionLongLoops : public ReductionLoops<X, Z, X> {
2019-06-06 14:21:15 +02:00
public:
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
static void wrapper(int opNum, sd::memory::Workspace* workspace, const X* x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int* dims, X* extraParams);
2019-06-06 14:21:15 +02:00
template <typename OpType>
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
static void innerloopReduce(sd::memory::Workspace* workspace, const X* x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int* dims, X* extraParams);
2019-06-06 14:21:15 +02:00
};
template <typename X>
class ND4J_EXPORT ReductionSameLoops : public ReductionLoops<X, X, X> {
2019-06-06 14:21:15 +02:00
public:
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
static void wrapper(int opNum, sd::memory::Workspace* workspace, const X* x, const Nd4jLong *xShapeInfo, X* z, const Nd4jLong *zShapeInfo, const int* dims, X* extraParams);
2019-06-06 14:21:15 +02:00
template <typename OpType>
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
static void innerloopReduce(sd::memory::Workspace* workspace, const X* x, const Nd4jLong *xShapeInfo, X* z, const Nd4jLong *zShapeInfo, const int* dims, X* extraParams);
2019-06-06 14:21:15 +02:00
};
template <typename X, typename Z>
2019-06-06 14:21:15 +02:00
class ND4J_EXPORT IndexReductionLoops {
private:
public:
Legacy API changes (#441) * initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored buffer() and shapeInfo() methods usage with NDArray class. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt Graph class methods to use const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt choose op to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt where op shape method to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt lstsq op to use constant empty shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt matrix_diag_part op shape routine to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt determinant ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt mean_pairwssqerr_loss ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for loss ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt log_loss op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt dilation2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted deconv2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted dynamicRNN op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for lstm layer ops. Signed-off-by: shugeo <sgazeos@gmail.com> * few updates Signed-off-by: raver119@gmail.com <raver119@gmail.com> * first cuda tweak Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Adopt constant shapes for sconv2d ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes for gru ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes with shape methods for segment ops and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with unsorted_segment_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with gamma op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods of reduce_stddev ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for reduce_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape method for squeeze op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt strided_slice shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored concat op shape method to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape method for mirror_pad op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted split op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted tile ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Added const cast for mkldnn routines handles. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored logSoftMaxForVector_ routine to conform with proper data and shape pointer casts. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetic changes to proper usage of constant pointers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple shape comparators for strides and addBias helpers to proper use data pointers with inplace option. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored depthToSpace helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored histogram helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored im2col helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored gather and gatherND helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage on percentile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed gather shape with helpers and range buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with space to depth helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage and constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with LUP decomposition> Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored onehot_ helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pad and prefix to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactoed softmax helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed space to batch helpers to use buffers properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed stack and split helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with sparse to dense helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with mindistance_ helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with tile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with legacy pairwise bool ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple of methods to adopt constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed broadcasting with constant shape." Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const usage with inplace reverse and constant shapes with legacy reduction. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored sort to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected sort for constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with special methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored Context to conform with constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * CUDA broadcasting headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * pairwise/indexreduce/random headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored native ops to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * legacy reduce3/scalar headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected pullRow signature and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected routines to proper use of constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with NDArray tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed native ops tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed special concat routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with test. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with a test. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored TAD.h and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored calcStrides* routines to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed miscelaneous errors with constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected definitions for declared functions. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed const shapes with shape routines. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed shape method for broadcastable case. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * xw_plus_b BP shape fn restored Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed signatures with broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Repaired backprops shape methods for a set of operations. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored broadcast bool for cuda. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods for 3 args with const qualifier. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed a couple of kernel signatures for broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels signatures for const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise methods to persistent buffers and shapes usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with scalar kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored indexreduce kernels signatures to use const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise bool kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored random special ops to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored native ops to conform with const shapes and buffers under cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetical changes only. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes and buffers error. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected start pos routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored helpers to use proper methods instead. Signed-off-by: shugeo <sgazeos@gmail.com> * bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected const shape cases with sort and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes for sort. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored kernel declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernel declarations to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed segment helpers kernels declarations and so on to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with segment and solve helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernel declaration with adjustWeight helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed cuda implementations for constant shape helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted const shape usage with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted top_k kernels to use const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernels declarations to adopt const shapes with helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored NDArray definitions to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes with image suppression helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Slight improvement with buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with definitions. Signed-off-by: shugeo <sgazeos@gmail.com> * minor updates on cpu side Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored const shape usage with ConstantDescritor and native ops with cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tear and tile kernels to adopt with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * softmax_loop fix Signed-off-by: raver119 <raver119@gmail.com> * update missing signature Signed-off-by: raver119@gmail.com <raver119@gmail.com> * softmax again Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more missing consts Signed-off-by: raver119 <raver119@gmail.com> * new methods updated Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: shugeo <sgazeos@gmail.com>
2020-05-09 07:06:14 +02:00
static void wrapIndexReduce(int opNum, const void* x, const Nd4jLong* xShapeInfo, void* z, const Nd4jLong* zShapeInfo, const Nd4jLong* tadShapeInfo, const Nd4jLong* tadOffsets, void* extraParams);
2019-06-06 14:21:15 +02:00
template <typename OpType>
Legacy API changes (#441) * initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored buffer() and shapeInfo() methods usage with NDArray class. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt Graph class methods to use const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt choose op to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt where op shape method to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt lstsq op to use constant empty shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt matrix_diag_part op shape routine to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt determinant ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt mean_pairwssqerr_loss ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for loss ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt log_loss op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt dilation2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted deconv2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted dynamicRNN op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for lstm layer ops. Signed-off-by: shugeo <sgazeos@gmail.com> * few updates Signed-off-by: raver119@gmail.com <raver119@gmail.com> * first cuda tweak Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Adopt constant shapes for sconv2d ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes for gru ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes with shape methods for segment ops and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with unsorted_segment_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with gamma op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods of reduce_stddev ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for reduce_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape method for squeeze op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt strided_slice shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored concat op shape method to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape method for mirror_pad op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted split op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted tile ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Added const cast for mkldnn routines handles. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored logSoftMaxForVector_ routine to conform with proper data and shape pointer casts. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetic changes to proper usage of constant pointers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple shape comparators for strides and addBias helpers to proper use data pointers with inplace option. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored depthToSpace helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored histogram helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored im2col helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored gather and gatherND helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage on percentile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed gather shape with helpers and range buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with space to depth helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage and constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with LUP decomposition> Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored onehot_ helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pad and prefix to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactoed softmax helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed space to batch helpers to use buffers properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed stack and split helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with sparse to dense helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with mindistance_ helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with tile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with legacy pairwise bool ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple of methods to adopt constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed broadcasting with constant shape." Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const usage with inplace reverse and constant shapes with legacy reduction. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored sort to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected sort for constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with special methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored Context to conform with constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * CUDA broadcasting headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * pairwise/indexreduce/random headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored native ops to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * legacy reduce3/scalar headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected pullRow signature and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected routines to proper use of constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with NDArray tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed native ops tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed special concat routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with test. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with a test. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored TAD.h and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored calcStrides* routines to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed miscelaneous errors with constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected definitions for declared functions. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed const shapes with shape routines. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed shape method for broadcastable case. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * xw_plus_b BP shape fn restored Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed signatures with broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Repaired backprops shape methods for a set of operations. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored broadcast bool for cuda. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods for 3 args with const qualifier. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed a couple of kernel signatures for broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels signatures for const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise methods to persistent buffers and shapes usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with scalar kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored indexreduce kernels signatures to use const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise bool kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored random special ops to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored native ops to conform with const shapes and buffers under cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetical changes only. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes and buffers error. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected start pos routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored helpers to use proper methods instead. Signed-off-by: shugeo <sgazeos@gmail.com> * bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected const shape cases with sort and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes for sort. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored kernel declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernel declarations to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed segment helpers kernels declarations and so on to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with segment and solve helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernel declaration with adjustWeight helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed cuda implementations for constant shape helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted const shape usage with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted top_k kernels to use const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernels declarations to adopt const shapes with helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored NDArray definitions to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes with image suppression helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Slight improvement with buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with definitions. Signed-off-by: shugeo <sgazeos@gmail.com> * minor updates on cpu side Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored const shape usage with ConstantDescritor and native ops with cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tear and tile kernels to adopt with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * softmax_loop fix Signed-off-by: raver119 <raver119@gmail.com> * update missing signature Signed-off-by: raver119@gmail.com <raver119@gmail.com> * softmax again Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more missing consts Signed-off-by: raver119 <raver119@gmail.com> * new methods updated Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: shugeo <sgazeos@gmail.com>
2020-05-09 07:06:14 +02:00
static void loopIndexReduce(const X* x, const Nd4jLong* xShapeInfo, Z* z, const Nd4jLong* zShapeInfo, const Nd4jLong* tadShapeInfo, const Nd4jLong* tadOffsets, X* extraParams);
2019-06-06 14:21:15 +02:00
};
template <typename X, typename Z, typename E>
class ND4J_EXPORT TransformLoops {
public:
template<typename OpType>
Legacy API changes (#441) * initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored buffer() and shapeInfo() methods usage with NDArray class. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt Graph class methods to use const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt choose op to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt where op shape method to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt lstsq op to use constant empty shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt matrix_diag_part op shape routine to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt determinant ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt mean_pairwssqerr_loss ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for loss ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt log_loss op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt dilation2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted deconv2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted dynamicRNN op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for lstm layer ops. Signed-off-by: shugeo <sgazeos@gmail.com> * few updates Signed-off-by: raver119@gmail.com <raver119@gmail.com> * first cuda tweak Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Adopt constant shapes for sconv2d ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes for gru ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes with shape methods for segment ops and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with unsorted_segment_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with gamma op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods of reduce_stddev ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for reduce_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape method for squeeze op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt strided_slice shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored concat op shape method to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape method for mirror_pad op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted split op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted tile ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Added const cast for mkldnn routines handles. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored logSoftMaxForVector_ routine to conform with proper data and shape pointer casts. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetic changes to proper usage of constant pointers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple shape comparators for strides and addBias helpers to proper use data pointers with inplace option. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored depthToSpace helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored histogram helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored im2col helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored gather and gatherND helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage on percentile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed gather shape with helpers and range buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with space to depth helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage and constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with LUP decomposition> Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored onehot_ helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pad and prefix to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactoed softmax helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed space to batch helpers to use buffers properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed stack and split helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with sparse to dense helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with mindistance_ helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with tile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with legacy pairwise bool ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple of methods to adopt constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed broadcasting with constant shape." Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const usage with inplace reverse and constant shapes with legacy reduction. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored sort to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected sort for constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with special methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored Context to conform with constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * CUDA broadcasting headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * pairwise/indexreduce/random headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored native ops to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * legacy reduce3/scalar headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected pullRow signature and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected routines to proper use of constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with NDArray tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed native ops tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed special concat routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with test. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with a test. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored TAD.h and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored calcStrides* routines to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed miscelaneous errors with constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected definitions for declared functions. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed const shapes with shape routines. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed shape method for broadcastable case. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * xw_plus_b BP shape fn restored Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed signatures with broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Repaired backprops shape methods for a set of operations. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored broadcast bool for cuda. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods for 3 args with const qualifier. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed a couple of kernel signatures for broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels signatures for const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise methods to persistent buffers and shapes usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with scalar kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored indexreduce kernels signatures to use const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise bool kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored random special ops to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored native ops to conform with const shapes and buffers under cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetical changes only. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes and buffers error. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected start pos routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored helpers to use proper methods instead. Signed-off-by: shugeo <sgazeos@gmail.com> * bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected const shape cases with sort and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes for sort. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored kernel declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernel declarations to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed segment helpers kernels declarations and so on to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with segment and solve helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernel declaration with adjustWeight helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed cuda implementations for constant shape helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted const shape usage with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted top_k kernels to use const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernels declarations to adopt const shapes with helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored NDArray definitions to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes with image suppression helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Slight improvement with buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with definitions. Signed-off-by: shugeo <sgazeos@gmail.com> * minor updates on cpu side Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored const shape usage with ConstantDescritor and native ops with cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tear and tile kernels to adopt with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * softmax_loop fix Signed-off-by: raver119 <raver119@gmail.com> * update missing signature Signed-off-by: raver119@gmail.com <raver119@gmail.com> * softmax again Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more missing consts Signed-off-by: raver119 <raver119@gmail.com> * new methods updated Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: shugeo <sgazeos@gmail.com>
2020-05-09 07:06:14 +02:00
static FORCEINLINE void loopTransform(const X* x, const Nd4jLong* xShapeInfo, Z* z, const Nd4jLong* zShapeInfo, E* extraParams, uint64_t threadId, uint64_t numThreads);
2019-06-06 14:21:15 +02:00
};
template <typename X, typename Z>
class ND4J_EXPORT Reduction3Loops {
public:
template <typename OpType>
Legacy API changes (#441) * initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored buffer() and shapeInfo() methods usage with NDArray class. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt Graph class methods to use const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt choose op to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt where op shape method to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt lstsq op to use constant empty shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt matrix_diag_part op shape routine to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt determinant ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt mean_pairwssqerr_loss ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for loss ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt log_loss op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt dilation2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted deconv2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted dynamicRNN op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for lstm layer ops. Signed-off-by: shugeo <sgazeos@gmail.com> * few updates Signed-off-by: raver119@gmail.com <raver119@gmail.com> * first cuda tweak Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Adopt constant shapes for sconv2d ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes for gru ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes with shape methods for segment ops and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with unsorted_segment_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with gamma op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods of reduce_stddev ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for reduce_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape method for squeeze op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt strided_slice shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored concat op shape method to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape method for mirror_pad op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted split op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted tile ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Added const cast for mkldnn routines handles. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored logSoftMaxForVector_ routine to conform with proper data and shape pointer casts. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetic changes to proper usage of constant pointers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple shape comparators for strides and addBias helpers to proper use data pointers with inplace option. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored depthToSpace helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored histogram helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored im2col helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored gather and gatherND helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage on percentile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed gather shape with helpers and range buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with space to depth helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage and constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with LUP decomposition> Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored onehot_ helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pad and prefix to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactoed softmax helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed space to batch helpers to use buffers properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed stack and split helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with sparse to dense helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with mindistance_ helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with tile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with legacy pairwise bool ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple of methods to adopt constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed broadcasting with constant shape." Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const usage with inplace reverse and constant shapes with legacy reduction. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored sort to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected sort for constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with special methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored Context to conform with constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * CUDA broadcasting headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * pairwise/indexreduce/random headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored native ops to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * legacy reduce3/scalar headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected pullRow signature and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected routines to proper use of constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with NDArray tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed native ops tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed special concat routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with test. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with a test. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored TAD.h and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored calcStrides* routines to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed miscelaneous errors with constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected definitions for declared functions. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed const shapes with shape routines. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed shape method for broadcastable case. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * xw_plus_b BP shape fn restored Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed signatures with broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Repaired backprops shape methods for a set of operations. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored broadcast bool for cuda. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods for 3 args with const qualifier. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed a couple of kernel signatures for broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels signatures for const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise methods to persistent buffers and shapes usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with scalar kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored indexreduce kernels signatures to use const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise bool kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored random special ops to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored native ops to conform with const shapes and buffers under cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetical changes only. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes and buffers error. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected start pos routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored helpers to use proper methods instead. Signed-off-by: shugeo <sgazeos@gmail.com> * bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected const shape cases with sort and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes for sort. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored kernel declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernel declarations to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed segment helpers kernels declarations and so on to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with segment and solve helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernel declaration with adjustWeight helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed cuda implementations for constant shape helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted const shape usage with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted top_k kernels to use const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernels declarations to adopt const shapes with helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored NDArray definitions to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes with image suppression helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Slight improvement with buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with definitions. Signed-off-by: shugeo <sgazeos@gmail.com> * minor updates on cpu side Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored const shape usage with ConstantDescritor and native ops with cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tear and tile kernels to adopt with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * softmax_loop fix Signed-off-by: raver119 <raver119@gmail.com> * update missing signature Signed-off-by: raver119@gmail.com <raver119@gmail.com> * softmax again Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more missing consts Signed-off-by: raver119 <raver119@gmail.com> * new methods updated Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: shugeo <sgazeos@gmail.com>
2020-05-09 07:06:14 +02:00
static FORCEINLINE void loopReduce3(const X* x, const Nd4jLong* xShapeInfo, const X* y, const Nd4jLong* yShapeInfo, Z* z, const Nd4jLong* zShapeInfo, int* dims, int dimsLen, Z* extraParams, int64_t start, int64_t stop);
2019-06-06 14:21:15 +02:00
template <typename OpType>
Legacy API changes (#441) * initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored buffer() and shapeInfo() methods usage with NDArray class. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt Graph class methods to use const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt choose op to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt where op shape method to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt lstsq op to use constant empty shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt matrix_diag_part op shape routine to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt determinant ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt mean_pairwssqerr_loss ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for loss ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt log_loss op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt dilation2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted deconv2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted dynamicRNN op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for lstm layer ops. Signed-off-by: shugeo <sgazeos@gmail.com> * few updates Signed-off-by: raver119@gmail.com <raver119@gmail.com> * first cuda tweak Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Adopt constant shapes for sconv2d ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes for gru ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes with shape methods for segment ops and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with unsorted_segment_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with gamma op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods of reduce_stddev ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for reduce_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape method for squeeze op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt strided_slice shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored concat op shape method to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape method for mirror_pad op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted split op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted tile ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Added const cast for mkldnn routines handles. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored logSoftMaxForVector_ routine to conform with proper data and shape pointer casts. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetic changes to proper usage of constant pointers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple shape comparators for strides and addBias helpers to proper use data pointers with inplace option. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored depthToSpace helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored histogram helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored im2col helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored gather and gatherND helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage on percentile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed gather shape with helpers and range buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with space to depth helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage and constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with LUP decomposition> Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored onehot_ helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pad and prefix to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactoed softmax helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed space to batch helpers to use buffers properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed stack and split helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with sparse to dense helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with mindistance_ helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with tile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with legacy pairwise bool ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple of methods to adopt constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed broadcasting with constant shape." Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const usage with inplace reverse and constant shapes with legacy reduction. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored sort to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected sort for constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with special methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored Context to conform with constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * CUDA broadcasting headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * pairwise/indexreduce/random headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored native ops to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * legacy reduce3/scalar headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected pullRow signature and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected routines to proper use of constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with NDArray tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed native ops tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed special concat routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with test. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with a test. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored TAD.h and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored calcStrides* routines to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed miscelaneous errors with constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected definitions for declared functions. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed const shapes with shape routines. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed shape method for broadcastable case. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * xw_plus_b BP shape fn restored Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed signatures with broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Repaired backprops shape methods for a set of operations. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored broadcast bool for cuda. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods for 3 args with const qualifier. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed a couple of kernel signatures for broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels signatures for const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise methods to persistent buffers and shapes usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with scalar kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored indexreduce kernels signatures to use const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise bool kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored random special ops to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored native ops to conform with const shapes and buffers under cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetical changes only. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes and buffers error. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected start pos routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored helpers to use proper methods instead. Signed-off-by: shugeo <sgazeos@gmail.com> * bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected const shape cases with sort and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes for sort. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored kernel declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernel declarations to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed segment helpers kernels declarations and so on to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with segment and solve helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernel declaration with adjustWeight helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed cuda implementations for constant shape helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted const shape usage with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted top_k kernels to use const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernels declarations to adopt const shapes with helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored NDArray definitions to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes with image suppression helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Slight improvement with buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with definitions. Signed-off-by: shugeo <sgazeos@gmail.com> * minor updates on cpu side Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored const shape usage with ConstantDescritor and native ops with cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tear and tile kernels to adopt with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * softmax_loop fix Signed-off-by: raver119 <raver119@gmail.com> * update missing signature Signed-off-by: raver119@gmail.com <raver119@gmail.com> * softmax again Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more missing consts Signed-off-by: raver119 <raver119@gmail.com> * new methods updated Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: shugeo <sgazeos@gmail.com>
2020-05-09 07:06:14 +02:00
static FORCEINLINE void loopReduce3All(const X* x, const Nd4jLong* xShapeInfo, const X* y, const Nd4jLong* yShapeInfo, Z* z, const Nd4jLong* zShapeInfo, const Nd4jLong* xTadShapeInfo, const Nd4jLong* xTadOffsets, const Nd4jLong* yTadShapeInfo, const Nd4jLong* yTadOffsets, Z* extraParams, int64_t start, int64_t stop);
2019-06-06 14:21:15 +02:00
Legacy API changes (#441) * initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored buffer() and shapeInfo() methods usage with NDArray class. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt Graph class methods to use const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt choose op to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt where op shape method to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt lstsq op to use constant empty shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt matrix_diag_part op shape routine to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt determinant ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt mean_pairwssqerr_loss ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for loss ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt log_loss op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt dilation2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted deconv2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted dynamicRNN op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for lstm layer ops. Signed-off-by: shugeo <sgazeos@gmail.com> * few updates Signed-off-by: raver119@gmail.com <raver119@gmail.com> * first cuda tweak Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Adopt constant shapes for sconv2d ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes for gru ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes with shape methods for segment ops and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with unsorted_segment_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with gamma op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods of reduce_stddev ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for reduce_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape method for squeeze op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt strided_slice shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored concat op shape method to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape method for mirror_pad op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted split op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted tile ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Added const cast for mkldnn routines handles. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored logSoftMaxForVector_ routine to conform with proper data and shape pointer casts. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetic changes to proper usage of constant pointers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple shape comparators for strides and addBias helpers to proper use data pointers with inplace option. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored depthToSpace helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored histogram helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored im2col helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored gather and gatherND helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage on percentile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed gather shape with helpers and range buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with space to depth helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage and constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with LUP decomposition> Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored onehot_ helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pad and prefix to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactoed softmax helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed space to batch helpers to use buffers properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed stack and split helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with sparse to dense helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with mindistance_ helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with tile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with legacy pairwise bool ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple of methods to adopt constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed broadcasting with constant shape." Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const usage with inplace reverse and constant shapes with legacy reduction. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored sort to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected sort for constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with special methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored Context to conform with constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * CUDA broadcasting headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * pairwise/indexreduce/random headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored native ops to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * legacy reduce3/scalar headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected pullRow signature and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected routines to proper use of constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with NDArray tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed native ops tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed special concat routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with test. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with a test. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored TAD.h and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored calcStrides* routines to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed miscelaneous errors with constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected definitions for declared functions. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed const shapes with shape routines. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed shape method for broadcastable case. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * xw_plus_b BP shape fn restored Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed signatures with broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Repaired backprops shape methods for a set of operations. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored broadcast bool for cuda. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods for 3 args with const qualifier. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed a couple of kernel signatures for broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels signatures for const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise methods to persistent buffers and shapes usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with scalar kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored indexreduce kernels signatures to use const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise bool kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored random special ops to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored native ops to conform with const shapes and buffers under cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetical changes only. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes and buffers error. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected start pos routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored helpers to use proper methods instead. Signed-off-by: shugeo <sgazeos@gmail.com> * bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected const shape cases with sort and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes for sort. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored kernel declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernel declarations to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed segment helpers kernels declarations and so on to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with segment and solve helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernel declaration with adjustWeight helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed cuda implementations for constant shape helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted const shape usage with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted top_k kernels to use const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernels declarations to adopt const shapes with helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored NDArray definitions to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes with image suppression helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Slight improvement with buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with definitions. Signed-off-by: shugeo <sgazeos@gmail.com> * minor updates on cpu side Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored const shape usage with ConstantDescritor and native ops with cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tear and tile kernels to adopt with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * softmax_loop fix Signed-off-by: raver119 <raver119@gmail.com> * update missing signature Signed-off-by: raver119@gmail.com <raver119@gmail.com> * softmax again Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more missing consts Signed-off-by: raver119 <raver119@gmail.com> * new methods updated Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: shugeo <sgazeos@gmail.com>
2020-05-09 07:06:14 +02:00
static void wrapper(int opNum, const X* x, const Nd4jLong* xShapeInfo, const X* y, const Nd4jLong* yShapeInfo, Z* z, const Nd4jLong* zShapeInfo, int* dims, int dimsLen, Z* extraParams, int64_t start, int64_t stop);
2019-06-06 14:21:15 +02:00
Legacy API changes (#441) * initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored buffer() and shapeInfo() methods usage with NDArray class. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt Graph class methods to use const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt choose op to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt where op shape method to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt lstsq op to use constant empty shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt matrix_diag_part op shape routine to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt determinant ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt mean_pairwssqerr_loss ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for loss ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt log_loss op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt dilation2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted deconv2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted dynamicRNN op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for lstm layer ops. Signed-off-by: shugeo <sgazeos@gmail.com> * few updates Signed-off-by: raver119@gmail.com <raver119@gmail.com> * first cuda tweak Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Adopt constant shapes for sconv2d ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes for gru ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes with shape methods for segment ops and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with unsorted_segment_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with gamma op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods of reduce_stddev ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for reduce_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape method for squeeze op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt strided_slice shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored concat op shape method to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape method for mirror_pad op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted split op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted tile ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Added const cast for mkldnn routines handles. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored logSoftMaxForVector_ routine to conform with proper data and shape pointer casts. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetic changes to proper usage of constant pointers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple shape comparators for strides and addBias helpers to proper use data pointers with inplace option. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored depthToSpace helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored histogram helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored im2col helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored gather and gatherND helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage on percentile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed gather shape with helpers and range buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with space to depth helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage and constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with LUP decomposition> Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored onehot_ helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pad and prefix to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactoed softmax helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed space to batch helpers to use buffers properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed stack and split helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with sparse to dense helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with mindistance_ helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with tile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with legacy pairwise bool ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple of methods to adopt constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed broadcasting with constant shape." Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const usage with inplace reverse and constant shapes with legacy reduction. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored sort to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected sort for constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with special methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored Context to conform with constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * CUDA broadcasting headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * pairwise/indexreduce/random headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored native ops to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * legacy reduce3/scalar headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected pullRow signature and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected routines to proper use of constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with NDArray tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed native ops tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed special concat routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with test. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with a test. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored TAD.h and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored calcStrides* routines to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed miscelaneous errors with constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected definitions for declared functions. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed const shapes with shape routines. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed shape method for broadcastable case. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * xw_plus_b BP shape fn restored Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed signatures with broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Repaired backprops shape methods for a set of operations. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored broadcast bool for cuda. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods for 3 args with const qualifier. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed a couple of kernel signatures for broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels signatures for const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise methods to persistent buffers and shapes usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with scalar kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored indexreduce kernels signatures to use const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise bool kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored random special ops to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored native ops to conform with const shapes and buffers under cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetical changes only. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes and buffers error. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected start pos routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored helpers to use proper methods instead. Signed-off-by: shugeo <sgazeos@gmail.com> * bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected const shape cases with sort and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes for sort. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored kernel declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernel declarations to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed segment helpers kernels declarations and so on to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with segment and solve helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernel declaration with adjustWeight helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed cuda implementations for constant shape helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted const shape usage with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted top_k kernels to use const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernels declarations to adopt const shapes with helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored NDArray definitions to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes with image suppression helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Slight improvement with buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with definitions. Signed-off-by: shugeo <sgazeos@gmail.com> * minor updates on cpu side Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored const shape usage with ConstantDescritor and native ops with cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tear and tile kernels to adopt with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * softmax_loop fix Signed-off-by: raver119 <raver119@gmail.com> * update missing signature Signed-off-by: raver119@gmail.com <raver119@gmail.com> * softmax again Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more missing consts Signed-off-by: raver119 <raver119@gmail.com> * new methods updated Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: shugeo <sgazeos@gmail.com>
2020-05-09 07:06:14 +02:00
static void wrapperAll(int opNum, const X* x, const Nd4jLong* xShapeInfo, const X* y, const Nd4jLong* yShapeInfo, Z* z, const Nd4jLong* zShapeInfo, const Nd4jLong* xTadShapeInfo, const Nd4jLong* xTadOffsets, const Nd4jLong* yTadShapeInfo, const Nd4jLong* yTadOffsets, Z* extraParams, int64_t start, int64_t stop);
2019-06-06 14:21:15 +02:00
template <typename OpType>
Legacy API changes (#441) * initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored buffer() and shapeInfo() methods usage with NDArray class. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt Graph class methods to use const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt choose op to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt where op shape method to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt lstsq op to use constant empty shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt matrix_diag_part op shape routine to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt determinant ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt mean_pairwssqerr_loss ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for loss ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt log_loss op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt dilation2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted deconv2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted dynamicRNN op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for lstm layer ops. Signed-off-by: shugeo <sgazeos@gmail.com> * few updates Signed-off-by: raver119@gmail.com <raver119@gmail.com> * first cuda tweak Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Adopt constant shapes for sconv2d ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes for gru ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes with shape methods for segment ops and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with unsorted_segment_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with gamma op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods of reduce_stddev ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for reduce_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape method for squeeze op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt strided_slice shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored concat op shape method to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape method for mirror_pad op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted split op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted tile ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Added const cast for mkldnn routines handles. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored logSoftMaxForVector_ routine to conform with proper data and shape pointer casts. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetic changes to proper usage of constant pointers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple shape comparators for strides and addBias helpers to proper use data pointers with inplace option. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored depthToSpace helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored histogram helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored im2col helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored gather and gatherND helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage on percentile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed gather shape with helpers and range buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with space to depth helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage and constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with LUP decomposition> Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored onehot_ helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pad and prefix to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactoed softmax helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed space to batch helpers to use buffers properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed stack and split helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with sparse to dense helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with mindistance_ helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with tile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with legacy pairwise bool ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple of methods to adopt constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed broadcasting with constant shape." Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const usage with inplace reverse and constant shapes with legacy reduction. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored sort to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected sort for constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with special methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored Context to conform with constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * CUDA broadcasting headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * pairwise/indexreduce/random headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored native ops to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * legacy reduce3/scalar headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected pullRow signature and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected routines to proper use of constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with NDArray tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed native ops tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed special concat routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with test. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with a test. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored TAD.h and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored calcStrides* routines to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed miscelaneous errors with constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected definitions for declared functions. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed const shapes with shape routines. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed shape method for broadcastable case. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * xw_plus_b BP shape fn restored Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed signatures with broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Repaired backprops shape methods for a set of operations. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored broadcast bool for cuda. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods for 3 args with const qualifier. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed a couple of kernel signatures for broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels signatures for const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise methods to persistent buffers and shapes usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with scalar kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored indexreduce kernels signatures to use const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise bool kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored random special ops to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored native ops to conform with const shapes and buffers under cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetical changes only. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes and buffers error. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected start pos routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored helpers to use proper methods instead. Signed-off-by: shugeo <sgazeos@gmail.com> * bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected const shape cases with sort and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes for sort. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored kernel declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernel declarations to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed segment helpers kernels declarations and so on to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with segment and solve helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernel declaration with adjustWeight helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed cuda implementations for constant shape helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted const shape usage with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted top_k kernels to use const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernels declarations to adopt const shapes with helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored NDArray definitions to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes with image suppression helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Slight improvement with buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with definitions. Signed-off-by: shugeo <sgazeos@gmail.com> * minor updates on cpu side Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored const shape usage with ConstantDescritor and native ops with cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tear and tile kernels to adopt with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * softmax_loop fix Signed-off-by: raver119 <raver119@gmail.com> * update missing signature Signed-off-by: raver119@gmail.com <raver119@gmail.com> * softmax again Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more missing consts Signed-off-by: raver119 <raver119@gmail.com> * new methods updated Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: shugeo <sgazeos@gmail.com>
2020-05-09 07:06:14 +02:00
static void innerloopReduce3(const X* x, const Nd4jLong* xShapeInfo, const X* y, const Nd4jLong* yShapeInfo, Z* z, const Nd4jLong* zShapeInfo, int* dims, int dimsLen, Z* extraParams, int64_t start, int64_t stop);
2019-06-06 14:21:15 +02:00
template <typename OpType>
Legacy API changes (#441) * initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored buffer() and shapeInfo() methods usage with NDArray class. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt Graph class methods to use const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt choose op to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt where op shape method to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt lstsq op to use constant empty shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt matrix_diag_part op shape routine to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt determinant ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt mean_pairwssqerr_loss ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for loss ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt log_loss op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt dilation2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted deconv2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted dynamicRNN op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for lstm layer ops. Signed-off-by: shugeo <sgazeos@gmail.com> * few updates Signed-off-by: raver119@gmail.com <raver119@gmail.com> * first cuda tweak Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Adopt constant shapes for sconv2d ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes for gru ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes with shape methods for segment ops and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with unsorted_segment_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with gamma op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods of reduce_stddev ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for reduce_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape method for squeeze op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt strided_slice shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored concat op shape method to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape method for mirror_pad op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted split op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted tile ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Added const cast for mkldnn routines handles. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored logSoftMaxForVector_ routine to conform with proper data and shape pointer casts. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetic changes to proper usage of constant pointers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple shape comparators for strides and addBias helpers to proper use data pointers with inplace option. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored depthToSpace helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored histogram helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored im2col helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored gather and gatherND helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage on percentile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed gather shape with helpers and range buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with space to depth helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage and constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with LUP decomposition> Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored onehot_ helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pad and prefix to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactoed softmax helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed space to batch helpers to use buffers properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed stack and split helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with sparse to dense helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with mindistance_ helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with tile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with legacy pairwise bool ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple of methods to adopt constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed broadcasting with constant shape." Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const usage with inplace reverse and constant shapes with legacy reduction. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored sort to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected sort for constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with special methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored Context to conform with constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * CUDA broadcasting headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * pairwise/indexreduce/random headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored native ops to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * legacy reduce3/scalar headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected pullRow signature and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected routines to proper use of constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with NDArray tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed native ops tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed special concat routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with test. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with a test. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored TAD.h and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored calcStrides* routines to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed miscelaneous errors with constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected definitions for declared functions. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed const shapes with shape routines. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed shape method for broadcastable case. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * xw_plus_b BP shape fn restored Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed signatures with broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Repaired backprops shape methods for a set of operations. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored broadcast bool for cuda. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods for 3 args with const qualifier. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed a couple of kernel signatures for broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels signatures for const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise methods to persistent buffers and shapes usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with scalar kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored indexreduce kernels signatures to use const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise bool kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored random special ops to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored native ops to conform with const shapes and buffers under cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetical changes only. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes and buffers error. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected start pos routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored helpers to use proper methods instead. Signed-off-by: shugeo <sgazeos@gmail.com> * bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected const shape cases with sort and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes for sort. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored kernel declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernel declarations to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed segment helpers kernels declarations and so on to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with segment and solve helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernel declaration with adjustWeight helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed cuda implementations for constant shape helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted const shape usage with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted top_k kernels to use const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernels declarations to adopt const shapes with helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored NDArray definitions to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes with image suppression helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Slight improvement with buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with definitions. Signed-off-by: shugeo <sgazeos@gmail.com> * minor updates on cpu side Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored const shape usage with ConstantDescritor and native ops with cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tear and tile kernels to adopt with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * softmax_loop fix Signed-off-by: raver119 <raver119@gmail.com> * update missing signature Signed-off-by: raver119@gmail.com <raver119@gmail.com> * softmax again Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more missing consts Signed-off-by: raver119 <raver119@gmail.com> * new methods updated Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: shugeo <sgazeos@gmail.com>
2020-05-09 07:06:14 +02:00
static void innerloopReduce3All(const X* x, const Nd4jLong* xShapeInfo, const X* y, const Nd4jLong* yShapeInfo, Z* z, const Nd4jLong* zShapeInfo, const Nd4jLong* xTadShapeInfo, const Nd4jLong* xTadOffsets, const Nd4jLong* yTadShapeInfo, const Nd4jLong* yTadOffsets, Z* extraParams, int64_t start, int64_t stop);
2019-06-06 14:21:15 +02:00
};
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
//////////////////////////////////////////////////////////////////////////
template <typename X, typename Z, typename E, typename OpType>
static void reduceExec21(const X *x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int *dims, E* extraParams) {
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const uint xAxis0 = shape::sizeAt(xShapeInfo, dims[0]);
const Nd4jLong xStrd0 = shape::strideAt(xShapeInfo, dims[0]);
const Nd4jLong zStrd0 = shape::strideAt(zShapeInfo, 0);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const uint xAxis1 = shape::sizeAt(xShapeInfo, dims[1]);
const Nd4jLong xStrd1 = shape::strideAt(xShapeInfo, dims[1]);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
auto func = PRAGMA_THREADS_FOR {
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
for (auto i0 = start; i0 < stop; ++i0) {
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
auto x0 = x + i0 * xStrd0;
auto z0 = z + i0 * zStrd0;
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
auto s = OpType::startingValue(x0);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
if(xStrd1 == 1)
for (uint i1 = 0; i1 < xAxis1; ++i1)
s = OpType::update(s, OpType::op(x0[i1], extraParams), extraParams);
else
for (uint i1 = 0; i1 < xAxis1; ++i1)
s = OpType::update(s, OpType::op(x0[i1 * xStrd1], extraParams), extraParams);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
*z0 = OpType::postProcess(s, static_cast<Nd4jLong>(xAxis1), extraParams);
}
};
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
samediff::Threads::parallel_for(func, 0,xAxis0);
}
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
//////////////////////////////////////////////////////////////////////////
template <typename X, typename Z, typename E, typename OpType>
static void reduceExec31(const X *x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int *dims, E* extraParams) {
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const uint xAxis0 = shape::sizeAt(xShapeInfo, dims[0]);
const Nd4jLong xStrd0 = shape::strideAt(xShapeInfo, dims[0]);
const Nd4jLong zStrd0 = shape::strideAt(zShapeInfo, 0);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const uint xAxis1 = shape::sizeAt(xShapeInfo, dims[1]);
const Nd4jLong xStrd1 = shape::strideAt(xShapeInfo, dims[1]);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const uint xAxis2 = shape::sizeAt(xShapeInfo, dims[2]);
const Nd4jLong xStrd2 = shape::strideAt(xShapeInfo, dims[2]);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const Nd4jLong tadLen = static_cast<Nd4jLong>(xAxis1 * xAxis2);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
auto func = PRAGMA_THREADS_FOR {
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
for (auto i0 = start; i0 < stop; ++i0) {
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
auto x0 = x + i0 * xStrd0;
auto z0 = z + i0 * zStrd0;
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
auto s = OpType::startingValue(x0);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
if(xStrd1 == 1)
for (uint i2 = 0; i2 < xAxis2; ++i2)
for (uint i1 = 0; i1 < xAxis1; ++i1)
s = OpType::update(s, OpType::op(x0[i1 + i2*xStrd2], extraParams), extraParams);
else if(xStrd2 == 1)
for (uint i1 = 0; i1 < xAxis1; ++i1)
for (uint i2 = 0; i2 < xAxis2; ++i2)
s = OpType::update(s, OpType::op(x0[i1*xStrd1 + i2], extraParams), extraParams);
else
for (uint i1 = 0; i1 < xAxis1; ++i1)
for (uint i2 = 0; i2 < xAxis2; ++i2)
s = OpType::update(s, OpType::op(x0[i1*xStrd1 + i2*xStrd2], extraParams), extraParams);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
*z0 = OpType::postProcess(s, tadLen, extraParams);
}
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
};
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
samediff::Threads::parallel_for(func, 0,xAxis0);
}
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
//////////////////////////////////////////////////////////////////////////
template <typename X, typename Z, typename E, typename OpType>
void reduceExec32(const X *x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int *dims, E* extraParams) {
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const uint xAxis0 = shape::sizeAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[0] : dims[1]);
const Nd4jLong xStrd0 = shape::strideAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[0] : dims[1]);
const Nd4jLong zStrd0 = shape::strideAt(zShapeInfo, shape::order(zShapeInfo) == 'c' ? 0 : 1);
const uint xAxis1 = shape::sizeAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[1] : dims[0]);
const Nd4jLong xStrd1 = shape::strideAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[1] : dims[0]);
const Nd4jLong zStrd1 = shape::strideAt(zShapeInfo, shape::order(zShapeInfo) == 'c' ? 1 : 0);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const uint xAxis2 = shape::sizeAt(xShapeInfo, dims[2]);
const Nd4jLong xStrd2 = shape::strideAt(xShapeInfo, dims[2]);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
auto func = PRAGMA_THREADS_FOR_2D {
for (auto i0 = start_x; i0 < stop_x; ++i0) {
for (auto i1 = start_y; i1 < stop_y; ++i1) {
auto x1 = x + i0 * xStrd0 + i1 * xStrd1;
auto z1 = z + i0 * zStrd0 + i1 * zStrd1;
auto s = OpType::startingValue(x1);
if(xStrd2 == 1)
for (uint i2 = 0; i2 < xAxis2; ++i2)
s = OpType::update(s, OpType::op(x1[i2], extraParams), extraParams);
else
for (uint i2 = 0; i2 < xAxis2; ++i2)
s = OpType::update(s, OpType::op(x1[i2 * xStrd2], extraParams), extraParams);
*z1 = OpType::postProcess(s, static_cast<Nd4jLong>(xAxis2), extraParams);
}
}
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
};
samediff::Threads::parallel_for(func, 0,xAxis0,1, 0,xAxis1,1);
}
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
//////////////////////////////////////////////////////////////////////////
template <typename X, typename Z, typename E, typename OpType>
void reduceExec41(const X *x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int *dims, E* extraParams) {
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const uint xAxis0 = shape::sizeAt(xShapeInfo, dims[0]);
const Nd4jLong xStrd0 = shape::strideAt(xShapeInfo, dims[0]);
const Nd4jLong zStrd0 = shape::strideAt(zShapeInfo, 0);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const uint xAxis1 = shape::sizeAt(xShapeInfo, dims[1]);
const Nd4jLong xStrd1 = shape::strideAt(xShapeInfo, dims[1]);
const uint xAxis2 = shape::sizeAt(xShapeInfo, dims[2]);
const Nd4jLong xStrd2 = shape::strideAt(xShapeInfo, dims[2]);
const uint xAxis3 = shape::sizeAt(xShapeInfo, dims[3]);
const Nd4jLong xStrd3 = shape::strideAt(xShapeInfo, dims[3]);
const Nd4jLong tadLen = static_cast<Nd4jLong>(xAxis1 * xAxis2 * xAxis3);
auto func = PRAGMA_THREADS_FOR {
for (auto i0 = start; i0 < stop; ++i0) {
auto x0 = x + i0 * xStrd0;
auto z0 = z + i0 * zStrd0;
auto s = OpType::startingValue(x0);
if(xStrd1 == 1)
for (uint i3 = 0; i3 < xAxis3; ++i3)
for (uint i2 = 0; i2 < xAxis2; ++i2)
for (uint i1 = 0; i1 < xAxis1; ++i1)
s = OpType::update(s, OpType::op(x0[i1 + i2*xStrd2 + i3*xStrd3], extraParams), extraParams);
else if(xStrd2 == 1)
for (uint i1 = 0; i1 < xAxis1; ++i1)
for (uint i3 = 0; i3 < xAxis3; ++i3)
for (uint i2 = 0; i2 < xAxis2; ++i2)
s = OpType::update(s, OpType::op(x0[i1*xStrd1 + i2 + i3*xStrd3], extraParams), extraParams);
else if(xStrd3 == 1)
for (uint i1 = 0; i1 < xAxis1; ++i1)
for (uint i2 = 0; i2 < xAxis2; ++i2)
for (uint i3 = 0; i3 < xAxis3; ++i3)
s = OpType::update(s, OpType::op(x0[i1*xStrd1 + i2*xStrd2 + i3], extraParams), extraParams);
else
for (uint i1 = 0; i1 < xAxis1; ++i1)
for (uint i2 = 0; i2 < xAxis2; ++i2)
for (uint i3 = 0; i3 < xAxis3; ++i3)
s = OpType::update(s, OpType::op(x0[i1*xStrd1 + i2*xStrd2 + i3*xStrd3], extraParams), extraParams);
*z0 = OpType::postProcess(s, tadLen, extraParams);
}
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
};
samediff::Threads::parallel_for(func, 0,xAxis0);
}
//////////////////////////////////////////////////////////////////////////
template <typename X, typename Z, typename E, typename OpType>
void reduceExec42(const X *x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int *dims, E* extraParams) {
const uint xAxis0 = shape::sizeAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[0] : dims[1]);
const Nd4jLong xStrd0 = shape::strideAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[0] : dims[1]);
const Nd4jLong zStrd0 = shape::strideAt(zShapeInfo, shape::order(zShapeInfo) == 'c' ? 0 : 1);
const uint xAxis1 = shape::sizeAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[1] : dims[0]);
const Nd4jLong xStrd1 = shape::strideAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[1] : dims[0]);
const Nd4jLong zStrd1 = shape::strideAt(zShapeInfo, shape::order(zShapeInfo) == 'c' ? 1 : 0);
const uint xAxis2 = shape::sizeAt(xShapeInfo, dims[2]);
const Nd4jLong xStrd2 = shape::strideAt(xShapeInfo, dims[2]);
const uint xAxis3 = shape::sizeAt(xShapeInfo, dims[3]);
const Nd4jLong xStrd3 = shape::strideAt(xShapeInfo, dims[3]);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const Nd4jLong tadLen = static_cast<Nd4jLong>(xAxis2 * xAxis3);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
auto func = PRAGMA_THREADS_FOR_2D {
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
for (auto i0 = start_x; i0 < stop_x; ++i0) {
for (auto i1 = start_y; i1 < stop_y; ++i1) {
auto x1 = x + i0 * xStrd0 + i1 * xStrd1;
auto z1 = z + i0 * zStrd0 + i1 * zStrd1;
auto s = OpType::startingValue(x1);
if(xStrd2 == 1)
for (uint i3 = 0; i3 < xAxis3; ++i3)
for (uint i2 = 0; i2 < xAxis2; ++i2)
s = OpType::update(s, OpType::op(x1[i2 + i3*xStrd3], extraParams), extraParams);
else if(xStrd3 == 1)
for (uint i2 = 0; i2 < xAxis2; ++i2)
for (uint i3 = 0; i3 < xAxis3; ++i3)
s = OpType::update(s, OpType::op(x1[i2*xStrd2 + i3], extraParams), extraParams);
else
for (uint i2 = 0; i2 < xAxis2; ++i2)
for (uint i3 = 0; i3 < xAxis3; ++i3)
s = OpType::update(s, OpType::op(x1[i2*xStrd2 + i3*xStrd3], extraParams), extraParams);
*z1 = OpType::postProcess(s, tadLen, extraParams);
}
}
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
};
samediff::Threads::parallel_for(func, 0,xAxis0,1, 0,xAxis1,1);
}
//////////////////////////////////////////////////////////////////////////
template <typename X, typename Z, typename E, typename OpType>
void reduceExec43(const X *x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int *dims, E* extraParams) {
const uint xAxis0 = shape::sizeAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[0] : dims[2]);
const Nd4jLong xStrd0 = shape::strideAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[0] : dims[2]);
const Nd4jLong zStrd0 = shape::strideAt(zShapeInfo, shape::order(zShapeInfo) == 'c' ? 0 : 2);
const uint xAxis1 = shape::sizeAt(xShapeInfo, dims[1]);
const Nd4jLong xStrd1 = shape::strideAt(xShapeInfo, dims[1]);
const Nd4jLong zStrd1 = shape::strideAt(zShapeInfo, 1);
const uint xAxis2 = shape::sizeAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[2] : dims[0]);
const Nd4jLong xStrd2 = shape::strideAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[2] : dims[0]);
const Nd4jLong zStrd2 = shape::strideAt(zShapeInfo, shape::order(zShapeInfo) == 'c' ? 2 : 0);
const uint xAxis3 = shape::sizeAt(xShapeInfo, dims[3]);
const Nd4jLong xStrd3 = shape::strideAt(xShapeInfo, dims[3]);
auto func = PRAGMA_THREADS_FOR_3D {
for (auto i0 = start_x; i0 < stop_x; ++i0) {
for (auto i1 = start_y; i1 < stop_y; ++i1) {
for (auto i2 = start_z; i2 < stop_z; ++i2) {
auto x2 = x + i0 * xStrd0 + i1 * xStrd1 + i2 * xStrd2;
auto z2 = z + i0 * zStrd0 + i1 * zStrd1 + i2 * zStrd2;
auto s = OpType::startingValue(x2);
if(xStrd3 == 1)
for (uint i3 = 0; i3 < xAxis3; ++i3)
s = OpType::update(s, OpType::op(x2[i3], extraParams), extraParams);
else
for (uint i3 = 0; i3 < xAxis3; ++i3)
s = OpType::update(s, OpType::op(x2[i3*xStrd3], extraParams), extraParams);
*z2 = OpType::postProcess(s, static_cast<Nd4jLong>(xAxis3), extraParams);
}
}
}
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
};
samediff::Threads::parallel_for(func, 0,xAxis0,1, 0,xAxis1,1, 0,xAxis2,1);
}
//////////////////////////////////////////////////////////////////////////
template <typename X, typename Z, typename E, typename OpType>
void reduceExec51(const X *x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int *dims, E* extraParams) {
const uint xAxis0 = shape::sizeAt(xShapeInfo, dims[0]);
const Nd4jLong xStrd0 = shape::strideAt(xShapeInfo, dims[0]);
const Nd4jLong zStrd0 = shape::strideAt(zShapeInfo, 0);
const uint xAxis1 = shape::sizeAt(xShapeInfo, dims[1]);
const Nd4jLong xStrd1 = shape::strideAt(xShapeInfo, dims[1]);
const uint xAxis2 = shape::sizeAt(xShapeInfo, dims[2]);
const Nd4jLong xStrd2 = shape::strideAt(xShapeInfo, dims[2]);
const uint xAxis3 = shape::sizeAt(xShapeInfo, dims[3]);
const Nd4jLong xStrd3 = shape::strideAt(xShapeInfo, dims[3]);
const uint xAxis4 = shape::sizeAt(xShapeInfo, dims[4]);
const Nd4jLong xStrd4 = shape::strideAt(xShapeInfo, dims[4]);
const Nd4jLong tadLen = static_cast<Nd4jLong>(xAxis1 * xAxis2 * xAxis3 * xAxis4);
auto func = PRAGMA_THREADS_FOR {
for (auto i0 = start; i0 < stop; ++i0) {
auto x0 = x + i0 * xStrd0;
auto z0 = z + i0 * zStrd0;
auto s = OpType::startingValue(x0);
if(xStrd1 == 1)
for (uint i4 = 0; i4 < xAxis4; ++i4)
for (uint i3 = 0; i3 < xAxis3; ++i3)
for (uint i2 = 0; i2 < xAxis2; ++i2)
for (uint i1 = 0; i1 < xAxis1; ++i1)
s = OpType::update(s, OpType::op(x0[i1 + i2*xStrd2 + i3*xStrd3 + i4*xStrd4], extraParams), extraParams);
else if(xStrd2 == 1)
for (uint i4 = 0; i4 < xAxis4; ++i4)
for (uint i3 = 0; i3 < xAxis3; ++i3)
for (uint i1 = 0; i1 < xAxis1; ++i1)
for (uint i2 = 0; i2 < xAxis2; ++i2)
s = OpType::update(s, OpType::op(x0[i1*xStrd1 + i2 + i3*xStrd3 + i4*xStrd4], extraParams), extraParams);
else if(xStrd3 == 1)
for (uint i1 = 0; i1 < xAxis1; ++i1)
for (uint i2 = 0; i2 < xAxis2; ++i2)
for (uint i4 = 0; i4 < xAxis4; ++i4)
for (uint i3 = 0; i3 < xAxis3; ++i3)
s = OpType::update(s, OpType::op(x0[i1*xStrd1 + i2*xStrd2 + i3 + i4*xStrd4], extraParams), extraParams);
else if(xStrd4 == 1)
for (uint i1 = 0; i1 < xAxis1; ++i1)
for (uint i2 = 0; i2 < xAxis2; ++i2)
for (uint i3 = 0; i3 < xAxis3; ++i3)
for (uint i4 = 0; i4 < xAxis4; ++i4)
s = OpType::update(s, OpType::op(x0[i1*xStrd1 + i2*xStrd2 + i3*xStrd3 + i4], extraParams), extraParams);
else
for (uint i1 = 0; i1 < xAxis1; ++i1)
for (uint i2 = 0; i2 < xAxis2; ++i2)
for (uint i3 = 0; i3 < xAxis3; ++i3)
for (uint i4 = 0; i4 < xAxis4; ++i4)
s = OpType::update(s, OpType::op(x0[i1*xStrd1 + i2*xStrd2 + i3*xStrd3 + i4*xStrd4], extraParams), extraParams);
*z0 = OpType::postProcess(s, tadLen, extraParams);
}
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
};
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
samediff::Threads::parallel_for(func, 0,xAxis0);
}
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
//////////////////////////////////////////////////////////////////////////
template <typename X, typename Z, typename E, typename OpType>
void reduceExec52(const X *x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int *dims, E* extraParams) {
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const uint xAxis0 = shape::sizeAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[0] : dims[1]);
const Nd4jLong xStrd0 = shape::strideAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[0] : dims[1]);
const Nd4jLong zStrd0 = shape::strideAt(zShapeInfo, shape::order(zShapeInfo) == 'c' ? 0 : 1);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const uint xAxis1 = shape::sizeAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[1] : dims[0]);
const Nd4jLong xStrd1 = shape::strideAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[1] : dims[0]);
const Nd4jLong zStrd1 = shape::strideAt(zShapeInfo, shape::order(zShapeInfo) == 'c' ? 1 : 0);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const uint xAxis2 = shape::sizeAt(xShapeInfo, dims[2]);
const Nd4jLong xStrd2 = shape::strideAt(xShapeInfo, dims[2]);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const uint xAxis3 = shape::sizeAt(xShapeInfo, dims[3]);
const Nd4jLong xStrd3 = shape::strideAt(xShapeInfo, dims[3]);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const uint xAxis4 = shape::sizeAt(xShapeInfo, dims[4]);
const Nd4jLong xStrd4 = shape::strideAt(xShapeInfo, dims[4]);
const Nd4jLong tadLen = static_cast<Nd4jLong>(xAxis2 * xAxis3 * xAxis4);
auto func = PRAGMA_THREADS_FOR_2D {
for (auto i0 = start_x; i0 < stop_x; ++i0) {
for (auto i1 = start_y; i1 < stop_y; ++i1) {
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
auto x1 = x + i0 * xStrd0 + i1 * xStrd1;
auto z1 = z + i0 * zStrd0 + i1 * zStrd1;
auto s = OpType::startingValue(x1);
if(xStrd2 == 1)
for (uint i4 = 0; i4 < xAxis4; ++i4)
for (uint i3 = 0; i3 < xAxis3; ++i3)
for (uint i2 = 0; i2 < xAxis2; ++i2)
s = OpType::update(s, OpType::op(x1[i2 + i3*xStrd3 + i4*xStrd4], extraParams), extraParams);
else if(xStrd3 == 1)
for (uint i2 = 0; i2 < xAxis2; ++i2)
for (uint i4 = 0; i4 < xAxis4; ++i4)
for (uint i3 = 0; i3 < xAxis3; ++i3)
s = OpType::update(s, OpType::op(x1[i2*xStrd2 + i3 + i4*xStrd4], extraParams), extraParams);
else if(xStrd4 == 1)
for (uint i2 = 0; i2 < xAxis2; ++i2)
for (uint i3 = 0; i3 < xAxis3; ++i3)
for (uint i4 = 0; i4 < xAxis4; ++i4)
s = OpType::update(s, OpType::op(x1[i2*xStrd2 + i3*xStrd3 + i4], extraParams), extraParams);
else
for (uint i2 = 0; i2 < xAxis2; ++i2)
for (uint i3 = 0; i3 < xAxis3; ++i3)
for (uint i4 = 0; i4 < xAxis4; ++i4)
s = OpType::update(s, OpType::op(x1[i2*xStrd2 + i3*xStrd3 + i4*xStrd4], extraParams), extraParams);
*z1 = OpType::postProcess(s, tadLen, extraParams);
}
}
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
};
samediff::Threads::parallel_for(func, 0,xAxis0,1, 0,xAxis1,1);
}
//////////////////////////////////////////////////////////////////////////
template <typename X, typename Z, typename E, typename OpType>
void reduceExec53(const X *x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int *dims, E* extraParams) {
const uint xAxis0 = shape::sizeAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[0] : dims[2]);
const Nd4jLong xStrd0 = shape::strideAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[0] : dims[2]);
const Nd4jLong zStrd0 = shape::strideAt(zShapeInfo, shape::order(zShapeInfo) == 'c' ? 0 : 2);
const uint xAxis1 = shape::sizeAt(xShapeInfo, dims[1]);
const Nd4jLong xStrd1 = shape::strideAt(xShapeInfo, dims[1]);
const Nd4jLong zStrd1 = shape::strideAt(zShapeInfo, 1);
const uint xAxis2 = shape::sizeAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[2] : dims[0]);
const Nd4jLong xStrd2 = shape::strideAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[2] : dims[0]);
const Nd4jLong zStrd2 = shape::strideAt(zShapeInfo, shape::order(zShapeInfo) == 'c' ? 2 : 0);
const uint xAxis3 = shape::sizeAt(xShapeInfo, dims[3]);
const Nd4jLong xStrd3 = shape::strideAt(xShapeInfo, dims[3]);
const uint xAxis4 = shape::sizeAt(xShapeInfo, dims[4]);
const Nd4jLong xStrd4 = shape::strideAt(xShapeInfo, dims[4]);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
const Nd4jLong tadLen = static_cast<Nd4jLong>(xAxis3 * xAxis4);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
auto func = PRAGMA_THREADS_FOR_3D {
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
for (auto i0 = start_x; i0 < stop_x; ++i0) {
for (auto i1 = start_y; i1 < stop_y; ++i1) {
for (auto i2 = start_z; i2 < stop_z; ++i2) {
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
auto x2 = x + i0 * xStrd0 + i1 * xStrd1 + i2 * xStrd2;
auto z2 = z + i0 * zStrd0 + i1 * zStrd1 + i2 * zStrd2;
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
auto s = OpType::startingValue(x2);
2019-06-06 14:21:15 +02:00
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
if(xStrd3 == 1)
for (uint i4 = 0; i4 < xAxis4; ++i4)
for (uint i3 = 0; i3 < xAxis3; ++i3)
s = OpType::update(s, OpType::op(x2[i3 + i4*xStrd4], extraParams), extraParams);
else if(xStrd4 == 1)
for (uint i3 = 0; i3 < xAxis3; ++i3)
for (uint i4 = 0; i4 < xAxis4; ++i4)
s = OpType::update(s, OpType::op(x2[i3*xStrd3 + i4], extraParams), extraParams);
else
for (uint i3 = 0; i3 < xAxis3; ++i3)
for (uint i4 = 0; i4 < xAxis4; ++i4)
s = OpType::update(s, OpType::op(x2[i3*xStrd3 + i4*xStrd4], extraParams), extraParams);
*z2 = OpType::postProcess(s, tadLen, extraParams);
}
}
}
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
};
samediff::Threads::parallel_for(func, 0,xAxis0,1, 0,xAxis1,1, 0,xAxis2,1);
}
//////////////////////////////////////////////////////////////////////////
template <typename X, typename Z, typename E, typename OpType>
void reduceExec54(const X *x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int *dims, E* extraParams) {
const uint xAxis0 = shape::sizeAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[0] : dims[3]);
const Nd4jLong xStrd0 = shape::strideAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[0] : dims[3]);
const Nd4jLong zStrd0 = shape::strideAt(zShapeInfo, shape::order(zShapeInfo) == 'c' ? 0 : 3);
const uint xAxis1 = shape::sizeAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[1] : dims[2]);
const Nd4jLong xStrd1 = shape::strideAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[1] : dims[2]);
const Nd4jLong zStrd1 = shape::strideAt(zShapeInfo, shape::order(zShapeInfo) == 'c' ? 1 : 2);
const uint xAxis2 = shape::sizeAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[2] : dims[1]);
const Nd4jLong xStrd2 = shape::strideAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[2] : dims[1]);
const Nd4jLong zStrd2 = shape::strideAt(zShapeInfo, shape::order(zShapeInfo) == 'c' ? 2 : 1);
const uint xAxis3 = shape::sizeAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[3] : dims[0]);
const Nd4jLong xStrd3 = shape::strideAt(xShapeInfo, shape::order(zShapeInfo) == 'c' ? dims[3] : dims[0]);
const Nd4jLong zStrd3 = shape::strideAt(zShapeInfo, shape::order(zShapeInfo) == 'c' ? 3 : 0);
const uint xAxis4 = shape::sizeAt(xShapeInfo, dims[4]);
const Nd4jLong xStrd4 = shape::strideAt(xShapeInfo, dims[4]);
auto func = PRAGMA_THREADS_FOR_3D {
for (auto i0 = start_x; i0 < stop_x; ++i0) {
for (auto i1 = start_y; i1 < stop_y; ++i1) {
for (auto i2 = start_z; i2 < stop_z; ++i2) {
for (auto i3 = 0; i3 < xAxis3; ++i3) {
auto x3 = x + i0 * xStrd0 + i1 * xStrd1 + i2 * xStrd2 + i3 * xStrd3;
auto z3 = z + i0 * zStrd0 + i1 * zStrd1 + i2 * zStrd2 + i3 * zStrd3;
auto s = OpType::startingValue(x3);
if(xStrd4 == 1)
for (uint i4 = 0; i4 < xAxis4; ++i4)
s = OpType::update(s, OpType::op(x3[i4], extraParams), extraParams);
else
for (uint i4 = 0; i4 < xAxis4; ++i4)
s = OpType::update(s, OpType::op(x3[i4*xStrd4], extraParams), extraParams);
*z3 = OpType::postProcess(s, static_cast<Nd4jLong>(xAxis4), extraParams);
}
}
}
2019-06-06 14:21:15 +02:00
}
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
};
samediff::Threads::parallel_for(func, 0,xAxis0,1, 0,xAxis1,1, 0,xAxis2,1);
}
////////////////////////////////////////////////////////////////////////
template <typename X, typename Z, typename E, typename OpType>
void reduceDefault(sd::memory::Workspace* workspace, const X *x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int *dims, E* extraParams) {
const int zRank = shape::rank(zShapeInfo);
const int tadRank = shape::rank(xShapeInfo) - zRank;
Nd4jLong* outerXTadShapeInfo = sd::ShapeBuilders::createSubArrShapeInfo(xShapeInfo, dims, zRank);
Nd4jLong* innerXTadShapeInfo = sd::ShapeBuilders::createSubArrShapeInfo(xShapeInfo, dims+zRank, tadRank);
const bool sameOffsets1 = shape::haveSameShapeAndStrides(zShapeInfo, outerXTadShapeInfo);
const bool sameOffsets2 = shape::haveSameShapeAndStrides(zShapeInfo, innerXTadShapeInfo);
const Nd4jLong zLen = shape::length(zShapeInfo);
const Nd4jLong tadLen = shape::length(innerXTadShapeInfo);
Nd4jLong* zOffsets = nullptr;
ALLOCATE(zOffsets, workspace, zLen, Nd4jLong);
shape::calcOffsets(zShapeInfo, zOffsets);
Nd4jLong* outerXTadOffsets = zOffsets;
if(!sameOffsets1) {
ALLOCATE(outerXTadOffsets, workspace, zLen, Nd4jLong);
shape::calcOffsets(outerXTadShapeInfo, outerXTadOffsets);
}
Nd4jLong* innerXTadOffsets = zOffsets;
if(!sameOffsets2) {
ALLOCATE(innerXTadOffsets, workspace, tadLen, Nd4jLong);
shape::calcOffsets(innerXTadShapeInfo, innerXTadOffsets);
2019-06-06 14:21:15 +02:00
}
Development updates (#9053) * RL4J: Add generic update rule (#502) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> * Shyrma reduce (#481) * - start working on improving of cpu legacy code for reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving legacy loops Signed-off-by: Yurii <iuriish@yahoo.com> * - still working on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on improving reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - testing speed run of new reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - working on improvement of default loop for reduce op Signed-off-by: Yurii <iuriish@yahoo.com> * - update signatures of stuff which calls reduce ops Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections in cuda reduce kernels Signed-off-by: Yurii <iuriish@yahoo.com> * - change loop for default case in broadcast legacy ops Signed-off-by: Yurii <iuriish@yahoo.com> * - comment some shape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - comment unnecessary prints in RNGtests Signed-off-by: Yurii <iuriish@yahoo.com> * - finish to resolve conflicts after master has been merged Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of some compilation mistakes of cuda stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor changes Signed-off-by: Yurii <iuriish@yahoo.com> * - further search for bug causing crash on java test Signed-off-by: Yurii <iuriish@yahoo.com> * - add scalar case in reduce_ ... exec stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - minor corrections in NAtiveOps.cu Signed-off-by: Yurii <iuriish@yahoo.com> * - add switch to scalar case execReduceXD functions Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in ConstantShapeHelper::createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> * - correct cuda mirrorPad Signed-off-by: Yurii <iuriish@yahoo.com> * - add support for vectors old shape in cuda createShapeInfoWithNoUnitiesForReduce Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> * Add support for CUDA 11.0 (#492) * Add support for CUDA 11.0 * libnd4j tweaks for CUDA 11 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * bindings update, again? Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update versions of JavaCPP Presets for FFmpeg, OpenBLAS, and NumPy * update API to match CUDA 8 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * * Update version of JavaCPP Presets for CPython * C++ updated for cuDNN 8.0 Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * 128-bit alignment for workspaces Signed-off-by: raver119@gmail.com <raver119@gmail.com> * change seed in 1 test Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fix dependecy duplication in python4j-parent pom * Fix group id for in python4j-numpy * few tests tweaked Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Remove macosx-x86_64-gpu from nd4j-tests-tensorflow * few minor tweaks for IndexReduce Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one test removed Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com> * RL4J: Add SyncTrainer and AgentLearnerBuilder for a few algorithms (#504) Signed-off-by: Alexandre Boulanger <aboulang2002@yahoo.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Yurii Shyrma <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com> Co-authored-by: Serhii Shepel <9946053+sshepel@users.noreply.github.com>
2020-07-26 14:59:27 +02:00
auto func = PRAGMA_THREADS_FOR{
for (auto i = start; i < stop; ++i) {
const auto tad = x + outerXTadOffsets[i];
auto s = OpType::startingValue(tad);
for (Nd4jLong j = 0; j < tadLen; j++)
s = OpType::update(s, OpType::op(tad[innerXTadOffsets[j]], extraParams), extraParams);
z[zOffsets[i]] = OpType::postProcess(s, tadLen, extraParams);
}
};
samediff::Threads::parallel_for(func, 0, shape::length(zShapeInfo));
RELEASE(outerXTadShapeInfo, workspace);
RELEASE(innerXTadShapeInfo, workspace);
RELEASE(zOffsets, workspace);
if(!sameOffsets1)
RELEASE(outerXTadOffsets, workspace);
if(!sameOffsets2)
RELEASE(innerXTadOffsets, workspace);
}
//////////////////////////////////////////////////////////////////////////////
template<typename X, typename Z, typename E>
template <typename OpType>
void sd::ReductionLoops<X, Z, E>::loopReduce(sd::memory::Workspace* workspace, const X* x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int* dims, E* extraParams) {
const int xRank = shape::rank(xShapeInfo);
const int zRank = shape::rank(zShapeInfo);
// shape::printShapeInfoLinear(xShapeInfo);
// shape::printShapeInfoLinear(zShapeInfo);
// shape::printIntArray(dims, shape::rank(xShapeInfo));
if(xRank == 2 && zRank == 1)
reduceExec21<X,Z,E,OpType>(x, xShapeInfo, z, zShapeInfo, dims, extraParams);
else if(xRank == 3 && zRank == 1)
reduceExec31<X,Z,E,OpType>(x, xShapeInfo, z, zShapeInfo, dims, extraParams);
else if(xRank == 3 && zRank == 2)
reduceExec32<X,Z,E,OpType>(x, xShapeInfo, z, zShapeInfo, dims, extraParams);
else if(xRank == 4 && zRank == 1)
reduceExec41<X,Z,E,OpType>(x, xShapeInfo, z, zShapeInfo, dims, extraParams);
else if(xRank == 4 && zRank == 2)
reduceExec42<X,Z,E,OpType>(x, xShapeInfo, z, zShapeInfo, dims, extraParams);
else if(xRank == 4 && zRank == 3)
reduceExec43<X,Z,E,OpType>(x, xShapeInfo, z, zShapeInfo, dims, extraParams);
else if(xRank == 5 && zRank == 1)
reduceExec51<X,Z,E,OpType>(x, xShapeInfo, z, zShapeInfo, dims, extraParams);
else if(xRank == 5 && zRank == 2)
reduceExec52<X,Z,E,OpType>(x, xShapeInfo, z, zShapeInfo, dims, extraParams);
else if(xRank == 5 && zRank == 3)
reduceExec53<X,Z,E,OpType>(x, xShapeInfo, z, zShapeInfo, dims, extraParams);
else if(xRank == 5 && zRank == 4)
reduceExec54<X,Z,E,OpType>(x, xShapeInfo, z, zShapeInfo, dims, extraParams);
else
reduceDefault<X,Z,E,OpType>(workspace, x, xShapeInfo, z, zShapeInfo, dims, extraParams);
}
2019-06-06 14:21:15 +02:00
//////////////////////////////////////////////////////////////////////////////
template <typename X, typename Z, typename E>
template <typename OpType>
Legacy API changes (#441) * initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored buffer() and shapeInfo() methods usage with NDArray class. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt Graph class methods to use const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt choose op to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt where op shape method to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt lstsq op to use constant empty shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt matrix_diag_part op shape routine to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt determinant ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt mean_pairwssqerr_loss ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for loss ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt log_loss op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt dilation2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted deconv2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted dynamicRNN op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for lstm layer ops. Signed-off-by: shugeo <sgazeos@gmail.com> * few updates Signed-off-by: raver119@gmail.com <raver119@gmail.com> * first cuda tweak Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Adopt constant shapes for sconv2d ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes for gru ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes with shape methods for segment ops and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with unsorted_segment_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with gamma op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods of reduce_stddev ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for reduce_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape method for squeeze op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt strided_slice shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored concat op shape method to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape method for mirror_pad op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted split op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted tile ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Added const cast for mkldnn routines handles. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored logSoftMaxForVector_ routine to conform with proper data and shape pointer casts. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetic changes to proper usage of constant pointers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple shape comparators for strides and addBias helpers to proper use data pointers with inplace option. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored depthToSpace helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored histogram helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored im2col helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored gather and gatherND helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage on percentile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed gather shape with helpers and range buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with space to depth helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage and constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with LUP decomposition> Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored onehot_ helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pad and prefix to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactoed softmax helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed space to batch helpers to use buffers properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed stack and split helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with sparse to dense helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with mindistance_ helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with tile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with legacy pairwise bool ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple of methods to adopt constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed broadcasting with constant shape." Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const usage with inplace reverse and constant shapes with legacy reduction. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored sort to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected sort for constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with special methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored Context to conform with constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * CUDA broadcasting headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * pairwise/indexreduce/random headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored native ops to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * legacy reduce3/scalar headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected pullRow signature and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected routines to proper use of constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with NDArray tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed native ops tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed special concat routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with test. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with a test. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored TAD.h and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored calcStrides* routines to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed miscelaneous errors with constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected definitions for declared functions. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed const shapes with shape routines. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed shape method for broadcastable case. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * xw_plus_b BP shape fn restored Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed signatures with broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Repaired backprops shape methods for a set of operations. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored broadcast bool for cuda. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods for 3 args with const qualifier. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed a couple of kernel signatures for broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels signatures for const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise methods to persistent buffers and shapes usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with scalar kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored indexreduce kernels signatures to use const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise bool kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored random special ops to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored native ops to conform with const shapes and buffers under cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetical changes only. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes and buffers error. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected start pos routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored helpers to use proper methods instead. Signed-off-by: shugeo <sgazeos@gmail.com> * bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected const shape cases with sort and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes for sort. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored kernel declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernel declarations to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed segment helpers kernels declarations and so on to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with segment and solve helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernel declaration with adjustWeight helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed cuda implementations for constant shape helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted const shape usage with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted top_k kernels to use const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernels declarations to adopt const shapes with helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored NDArray definitions to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes with image suppression helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Slight improvement with buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with definitions. Signed-off-by: shugeo <sgazeos@gmail.com> * minor updates on cpu side Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored const shape usage with ConstantDescritor and native ops with cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tear and tile kernels to adopt with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * softmax_loop fix Signed-off-by: raver119 <raver119@gmail.com> * update missing signature Signed-off-by: raver119@gmail.com <raver119@gmail.com> * softmax again Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more missing consts Signed-off-by: raver119 <raver119@gmail.com> * new methods updated Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: shugeo <sgazeos@gmail.com>
2020-05-09 07:06:14 +02:00
void sd::TransformLoops<X, Z, E>::loopTransform(const X* x, const Nd4jLong* xShapeInfo,
Z* z, const Nd4jLong* zShapeInfo,
E* extraParams,
uint64_t threadId, uint64_t numThreads) {
2019-06-06 14:21:15 +02:00
const LoopKind::Kind kindOfLoop = LoopKind::deduceKindOfLoopXZ(xShapeInfo, zShapeInfo);
const Nd4jLong* xShape = shape::shapeOf(const_cast<Nd4jLong*>(xShapeInfo));
2019-06-06 14:21:15 +02:00
const Nd4jLong* xStride = shape::stride(const_cast<Nd4jLong*>(xShapeInfo));
const Nd4jLong* zStride = shape::stride(const_cast<Nd4jLong*>(zShapeInfo));
const Nd4jLong len = shape::length(xShapeInfo);
if (len == 0)
return;
2019-06-06 14:21:15 +02:00
switch (kindOfLoop) {
//*********************************************//
case LoopKind::EWS1: {
auto span = samediff::Span::build(threadId, numThreads, 0, len, 1);
int64_t start = span.startX(), stop = span.stopX();
2019-06-06 14:21:15 +02:00
for (auto i = start; i < stop; i++)
z[i] = OpType::op(x[i], extraParams);
}
break;
2019-06-06 14:21:15 +02:00
//*********************************************//
case LoopKind::EWSNONZERO: {
const uint xEws = shape::elementWiseStride(xShapeInfo);
const uint zEws = shape::elementWiseStride(zShapeInfo);
2019-06-06 14:21:15 +02:00
auto span = samediff::Span::build(threadId, numThreads, 0, len, 1);
int64_t start = span.startX(), stop = span.stopX();
2019-06-06 14:21:15 +02:00
for (auto i = start; i < stop; i++)
z[i * zEws] = OpType::op(x[i * xEws], extraParams);
}
break;
2019-06-06 14:21:15 +02:00
//*********************************************//
case LoopKind::Z_EWSNONZERO: {
const uint zEws = shape::elementWiseStride(zShapeInfo);
uint castXShapeInfo[MAX_RANK];
const bool canCastX = sd::DataTypeUtils::castShapeInfo<uint>(xShapeInfo, castXShapeInfo);
2019-06-06 14:21:15 +02:00
auto span = samediff::Span::build(threadId, numThreads, 0, len, 1);
int64_t start = span.startX(), stop = span.stopX();
2019-06-06 14:21:15 +02:00
if (zEws > 1) {
for (auto i = start; i < stop; i++) {
const auto xOffset = shape::indexOffset(i, xShapeInfo, castXShapeInfo, canCastX);
z[i * zEws] = OpType::op(x[xOffset], extraParams);
2019-06-06 14:21:15 +02:00
}
}
else {
for (auto i = start; i < stop; i++) {
const auto xOffset = shape::indexOffset(i, xShapeInfo, castXShapeInfo, canCastX);
z[i] = OpType::op(x[xOffset], extraParams);
}
}
}
break;
2019-06-06 14:21:15 +02:00
//*********************************************//
case LoopKind::RANK1: {
auto span = samediff::Span::build(threadId, numThreads, 0, len, 1);
for (auto i0 = span.startX(); i0 < span.stopX(); i0++)
z[i0 * zStride[0]] = OpType::op(x[i0 * xStride[0]], extraParams);
}
break;
2019-06-06 14:21:15 +02:00
//*********************************************//
case LoopKind::RANK2: {
auto uXShape0 = static_cast<uint>(xShape[0]);
auto uXShape1 = static_cast<uint>(xShape[1]);
2019-06-06 14:21:15 +02:00
auto loop = samediff::ThreadsHelper::pickLoop2d(numThreads, uXShape0, uXShape1);
auto span = samediff::Span2::build(loop, threadId, numThreads, 0, uXShape0, 1, 0, uXShape1, 1);
2019-06-06 14:21:15 +02:00
for (auto i0 = span.startX(); i0 < span.stopX(); i0++) {
auto z0 = i0 * zStride[0];
auto x0 = i0 * xStride[0];
for (auto i1 = span.startY(); i1 < span.stopY(); ++i1)
z[z0 + i1 * zStride[1]] = OpType::op(x[x0 + i1 * xStride[1]], extraParams);
}
}
break;
2019-06-06 14:21:15 +02:00
//*********************************************//
case LoopKind::RANK3: {
auto uXShape0 = xShape[0];
auto uXShape1 = xShape[1];
auto uXShape2 = xShape[2];
2019-06-06 14:21:15 +02:00
auto loop = samediff::ThreadsHelper::pickLoop2d(numThreads, uXShape0, uXShape1);
auto span = samediff::Span2::build(loop, threadId, numThreads, 0, uXShape0, 1, 0, uXShape1, 1);
2019-06-06 14:21:15 +02:00
for (auto i0 = span.startX(); i0 < span.stopX(); i0++)
for (auto i1 = span.startY(); i1 < span.stopY(); i1++) {
auto z0 = i0 * zStride[0] + i1 * zStride[1];
auto x0 = i0 * xStride[0] + i1 * xStride[1];
for (Nd4jLong i2 = 0; i2 < uXShape2; ++i2)
z[z0 + i2 * zStride[2]] = OpType::op(x[x0 + i2 * xStride[2]], extraParams);
}
}
break;
2019-06-06 14:21:15 +02:00
//*********************************************//
case LoopKind::RANK4: {
auto uXShape0 = xShape[0];
auto uXShape1 = xShape[1];
auto uXShape2 = xShape[2];
auto uXShape3 = xShape[3];
auto loop = samediff::ThreadsHelper::pickLoop3d(numThreads, uXShape0, uXShape1, uXShape2);
auto span = samediff::Span3::build(loop, threadId, numThreads, 0, uXShape0, 1, 0, uXShape1, 1, 0, uXShape2, 1);
for (auto i0 = span.startX(); i0 < span.stopX(); i0++)
for (auto i1 = span.startY(); i1 < span.stopY(); i1++)
for (auto i2 = span.startZ(); i2 < span.stopZ(); i2++) {
auto x0 = i0 * xStride[0] + i1 * xStride[1] + i2 * xStride[2];
auto z0 = i0 * zStride[0] + i1 * zStride[1] + i2 * zStride[2];
for (Nd4jLong i3 = 0; i3 < uXShape3; ++i3)
z[z0 + i3 * zStride[3]] = OpType::op(x[x0 + i3 * xStride[3]], extraParams);
}
}
break;
2019-06-06 14:21:15 +02:00
//*********************************************//
case LoopKind::RANK5: {
auto uXShape0 = xShape[0];
auto uXShape1 = xShape[1];
auto uXShape2 = xShape[2];
auto uXShape3 = xShape[3];
auto uXShape4 = xShape[4];
auto loop = samediff::ThreadsHelper::pickLoop3d(numThreads, uXShape0, uXShape1, uXShape2);
auto span = samediff::Span3::build(loop, threadId, numThreads, 0, uXShape0, 1, 0, uXShape1, 1, 0, uXShape2, 1);
2019-06-06 14:21:15 +02:00
for (auto i0 = span.startX(); i0 < span.stopX(); i0++)
for (auto i1 = span.startY(); i1 < span.stopY(); i1++)
for (auto i2 = span.startZ(); i2 < span.stopZ(); i2++) {
auto z0 = i0 * zStride[0] + i1 * zStride[1] + i2 * zStride[2];
auto x0 = i0 * xStride[0] + i1 * xStride[1] + i2 * xStride[2];
2019-06-06 14:21:15 +02:00
for (Nd4jLong i3 = 0; i3 < uXShape3; ++i3) {
2019-06-06 14:21:15 +02:00
auto z1 = z0 + i3 * zStride[3];
auto x1 = x0 + i3 * xStride[3];
2019-06-06 14:21:15 +02:00
for (Nd4jLong i4 = 0; i4 < uXShape4; ++i4)
z[z1 + i4 * zStride[4]] = OpType::op(x[x1 + i4 * xStride[4]], extraParams);
2019-06-06 14:21:15 +02:00
}
}
}
break;
2019-06-06 14:21:15 +02:00
//*********************************************//
default: {
uint xShapeInfoCast[MAX_RANK];
uint zShapeInfoCast[MAX_RANK];
bool canCastX = DataTypeUtils::castShapeInfo(xShapeInfo, xShapeInfoCast);
bool canCastZ = DataTypeUtils::castShapeInfo(zShapeInfo, zShapeInfoCast);
auto span = samediff::Span::build(threadId, numThreads, 0, len, 1);
for (auto i = span.startX(); i < span.stopX(); i++) {
auto xOffset = shape::indexOffset(i, xShapeInfo, xShapeInfoCast, canCastX);
auto zOffset = shape::indexOffset(i, zShapeInfo, zShapeInfoCast, canCastZ);
z[zOffset] = OpType::op(x[xOffset], extraParams);
}
}
2019-06-06 14:21:15 +02:00
}
}
//////////////////////////////////////////////////////////////////////////////
2019-06-06 14:21:15 +02:00
template<typename X, typename Z>
template <typename OpType>
Legacy API changes (#441) * initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored buffer() and shapeInfo() methods usage with NDArray class. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt Graph class methods to use const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt choose op to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt where op shape method to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt lstsq op to use constant empty shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt matrix_diag_part op shape routine to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt determinant ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt mean_pairwssqerr_loss ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for loss ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt log_loss op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt dilation2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted deconv2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted dynamicRNN op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for lstm layer ops. Signed-off-by: shugeo <sgazeos@gmail.com> * few updates Signed-off-by: raver119@gmail.com <raver119@gmail.com> * first cuda tweak Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Adopt constant shapes for sconv2d ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes for gru ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes with shape methods for segment ops and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with unsorted_segment_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with gamma op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods of reduce_stddev ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for reduce_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape method for squeeze op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt strided_slice shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored concat op shape method to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape method for mirror_pad op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted split op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted tile ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Added const cast for mkldnn routines handles. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored logSoftMaxForVector_ routine to conform with proper data and shape pointer casts. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetic changes to proper usage of constant pointers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple shape comparators for strides and addBias helpers to proper use data pointers with inplace option. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored depthToSpace helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored histogram helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored im2col helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored gather and gatherND helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage on percentile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed gather shape with helpers and range buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with space to depth helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage and constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with LUP decomposition> Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored onehot_ helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pad and prefix to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactoed softmax helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed space to batch helpers to use buffers properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed stack and split helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with sparse to dense helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with mindistance_ helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with tile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with legacy pairwise bool ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple of methods to adopt constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed broadcasting with constant shape." Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const usage with inplace reverse and constant shapes with legacy reduction. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored sort to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected sort for constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with special methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored Context to conform with constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * CUDA broadcasting headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * pairwise/indexreduce/random headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored native ops to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * legacy reduce3/scalar headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected pullRow signature and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected routines to proper use of constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with NDArray tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed native ops tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed special concat routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with test. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with a test. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored TAD.h and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored calcStrides* routines to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed miscelaneous errors with constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected definitions for declared functions. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed const shapes with shape routines. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed shape method for broadcastable case. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * xw_plus_b BP shape fn restored Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed signatures with broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Repaired backprops shape methods for a set of operations. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored broadcast bool for cuda. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods for 3 args with const qualifier. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed a couple of kernel signatures for broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels signatures for const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise methods to persistent buffers and shapes usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with scalar kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored indexreduce kernels signatures to use const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise bool kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored random special ops to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored native ops to conform with const shapes and buffers under cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetical changes only. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes and buffers error. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected start pos routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored helpers to use proper methods instead. Signed-off-by: shugeo <sgazeos@gmail.com> * bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected const shape cases with sort and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes for sort. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored kernel declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernel declarations to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed segment helpers kernels declarations and so on to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with segment and solve helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernel declaration with adjustWeight helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed cuda implementations for constant shape helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted const shape usage with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted top_k kernels to use const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernels declarations to adopt const shapes with helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored NDArray definitions to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes with image suppression helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Slight improvement with buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with definitions. Signed-off-by: shugeo <sgazeos@gmail.com> * minor updates on cpu side Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored const shape usage with ConstantDescritor and native ops with cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tear and tile kernels to adopt with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * softmax_loop fix Signed-off-by: raver119 <raver119@gmail.com> * update missing signature Signed-off-by: raver119@gmail.com <raver119@gmail.com> * softmax again Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more missing consts Signed-off-by: raver119 <raver119@gmail.com> * new methods updated Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: shugeo <sgazeos@gmail.com>
2020-05-09 07:06:14 +02:00
void sd::Reduction3Loops<X, Z>::loopReduce3(const X* x, const Nd4jLong* xShapeInfo,
const X* y, const Nd4jLong* yShapeInfo,
Z* z, const Nd4jLong* zShapeInfo,
int* dims, int dimsLen,
Z* extraParameters, int64_t start, int64_t stop) {
2019-06-06 14:21:15 +02:00
// both tads have same shape, however strides and ews may differ
Z param0(OpType::startingValue(x)), param1(OpType::startingValue(x)), param2(extraParameters ? extraParameters[0] : OpType::startingValue(x));
const Nd4jLong xLen = shape::length(xShapeInfo);
const Nd4jLong yLen = shape::length(yShapeInfo);
Legacy API changes (#441) * initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored buffer() and shapeInfo() methods usage with NDArray class. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt Graph class methods to use const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt choose op to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt where op shape method to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt lstsq op to use constant empty shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt matrix_diag_part op shape routine to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt determinant ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt mean_pairwssqerr_loss ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for loss ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt log_loss op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt dilation2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted deconv2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted dynamicRNN op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for lstm layer ops. Signed-off-by: shugeo <sgazeos@gmail.com> * few updates Signed-off-by: raver119@gmail.com <raver119@gmail.com> * first cuda tweak Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Adopt constant shapes for sconv2d ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes for gru ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes with shape methods for segment ops and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with unsorted_segment_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with gamma op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods of reduce_stddev ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for reduce_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape method for squeeze op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt strided_slice shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored concat op shape method to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape method for mirror_pad op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted split op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted tile ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Added const cast for mkldnn routines handles. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored logSoftMaxForVector_ routine to conform with proper data and shape pointer casts. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetic changes to proper usage of constant pointers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple shape comparators for strides and addBias helpers to proper use data pointers with inplace option. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored depthToSpace helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored histogram helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored im2col helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored gather and gatherND helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage on percentile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed gather shape with helpers and range buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with space to depth helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage and constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with LUP decomposition> Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored onehot_ helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pad and prefix to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactoed softmax helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed space to batch helpers to use buffers properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed stack and split helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with sparse to dense helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with mindistance_ helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with tile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with legacy pairwise bool ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple of methods to adopt constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed broadcasting with constant shape." Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const usage with inplace reverse and constant shapes with legacy reduction. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored sort to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected sort for constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with special methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored Context to conform with constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * CUDA broadcasting headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * pairwise/indexreduce/random headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored native ops to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * legacy reduce3/scalar headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected pullRow signature and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected routines to proper use of constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with NDArray tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed native ops tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed special concat routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with test. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with a test. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored TAD.h and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored calcStrides* routines to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed miscelaneous errors with constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected definitions for declared functions. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed const shapes with shape routines. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed shape method for broadcastable case. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * xw_plus_b BP shape fn restored Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed signatures with broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Repaired backprops shape methods for a set of operations. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored broadcast bool for cuda. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods for 3 args with const qualifier. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed a couple of kernel signatures for broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels signatures for const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise methods to persistent buffers and shapes usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with scalar kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored indexreduce kernels signatures to use const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise bool kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored random special ops to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored native ops to conform with const shapes and buffers under cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetical changes only. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes and buffers error. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected start pos routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored helpers to use proper methods instead. Signed-off-by: shugeo <sgazeos@gmail.com> * bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected const shape cases with sort and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes for sort. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored kernel declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernel declarations to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed segment helpers kernels declarations and so on to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with segment and solve helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernel declaration with adjustWeight helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed cuda implementations for constant shape helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted const shape usage with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted top_k kernels to use const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernels declarations to adopt const shapes with helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored NDArray definitions to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes with image suppression helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Slight improvement with buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with definitions. Signed-off-by: shugeo <sgazeos@gmail.com> * minor updates on cpu side Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored const shape usage with ConstantDescritor and native ops with cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tear and tile kernels to adopt with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * softmax_loop fix Signed-off-by: raver119 <raver119@gmail.com> * update missing signature Signed-off-by: raver119@gmail.com <raver119@gmail.com> * softmax again Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more missing consts Signed-off-by: raver119 <raver119@gmail.com> * new methods updated Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: shugeo <sgazeos@gmail.com>
2020-05-09 07:06:14 +02:00
const Nd4jLong* xTadShapeInfo = nullptr, * yTadShapeInfo = nullptr, * xTadOffsets = nullptr, * yTadOffsets = nullptr;
2019-06-06 14:21:15 +02:00
TadPack tadPackX, tadPackY;
std::vector<Nd4jLong> zeroOffsets;
if (xLen == yLen) {
tadPackX = sd::ConstantTadHelper::getInstance().tadForDimensions(xShapeInfo, dims, dimsLen);
tadPackY = sd::ConstantTadHelper::getInstance().tadForDimensions(yShapeInfo, dims, dimsLen);
2019-06-06 14:21:15 +02:00
xTadShapeInfo = tadPackX.primaryShapeInfo();
yTadShapeInfo = tadPackY.primaryShapeInfo();
xTadOffsets = tadPackX.primaryOffsets();
yTadOffsets = tadPackY.primaryOffsets();
2019-06-06 14:21:15 +02:00
}
else if (yLen > xLen) {
tadPackY = sd::ConstantTadHelper::getInstance().tadForDimensions(yShapeInfo, dims, dimsLen);
2019-06-06 14:21:15 +02:00
xTadShapeInfo = xShapeInfo;
yTadShapeInfo = tadPackY.primaryShapeInfo();
yTadOffsets = tadPackY.primaryOffsets();
2019-06-06 14:21:15 +02:00
}
else {
tadPackX = sd::ConstantTadHelper::getInstance().tadForDimensions(xShapeInfo, dims, dimsLen);
2019-06-06 14:21:15 +02:00
yTadShapeInfo = yShapeInfo;
xTadShapeInfo = tadPackX.primaryShapeInfo();
xTadOffsets = tadPackX.primaryOffsets();
2019-06-06 14:21:15 +02:00
}
const LoopKind::Kind kindOfLoop = LoopKind::deduceKindOfLoopTadXYZ(xTadShapeInfo, yTadShapeInfo, zShapeInfo);
const auto xTadEws = shape::elementWiseStride(xTadShapeInfo);
const auto yTadEws = shape::elementWiseStride(yTadShapeInfo);
const auto zEws = shape::elementWiseStride(zShapeInfo);
2019-06-06 14:21:15 +02:00
const auto zLen = shape::length(zShapeInfo);
2019-06-06 14:21:15 +02:00
const auto tadLen = shape::length(xTadShapeInfo);
const auto tadShape = shape::shapeOf(xTadShapeInfo);
const auto xTadStride = shape::stride(xTadShapeInfo);
const auto yTadStride = shape::stride(xTadShapeInfo);
2019-06-06 14:21:15 +02:00
int numThreads = OmpLaunchHelper::tadThreads(tadLen, zLen);
switch (kindOfLoop) {
//*********************************************//
case LoopKind::EWS1: {
Z extraParams[3];
for (auto i = start; i < stop; i++) {
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = xTadOffsets ? x + xTadOffsets[i] : x;
const auto yTad = yTadOffsets ? y + yTadOffsets[i] : y;
auto s = OpType::startingValue(xTad);
for (Nd4jLong j = 0; j < tadLen; ++j)
s = OpType::update(s, OpType::op(xTad[j], yTad[j], extraParams), extraParams);
z[i] = OpType::postProcess(s, tadLen, extraParams);
};
}
break;
//*********************************************//
case LoopKind::EWSNONZERO: {
Z extraParams[3];
for (auto i = start; i < stop; i++) {
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = xTadOffsets ? x + xTadOffsets[i] : x;
const auto yTad = yTadOffsets ? y + yTadOffsets[i] : y;
auto s = OpType::startingValue(xTad);
for (Nd4jLong j = 0; j < tadLen; ++j)
s = OpType::update(s, OpType::op(xTad[j * xTadEws], yTad[j * yTadEws], extraParams), extraParams);
z[i * zEws] = OpType::postProcess(s, tadLen, extraParams);
};
}
break;
//*********************************************//
case LoopKind::RANK1: {
Z extraParams[3];
for (auto i = start; i < stop; i++) {
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = xTadOffsets ? x + xTadOffsets[i] : x;
const auto yTad = yTadOffsets ? y + yTadOffsets[i] : y;
auto s = OpType::startingValue(xTad);
for (Nd4jLong i0 = 0; i0 < tadLen; ++i0) {
const auto xTadOffset = i0 * xTadStride[0];
const auto yTadOffset = i0 * yTadStride[0];
s = OpType::update(s, OpType::op(xTad[xTadOffset], yTad[yTadOffset], extraParams), extraParams);
}
z[i * zEws] = OpType::postProcess(s, tadLen, extraParams);
};
}
break;
//*********************************************//
case LoopKind::RANK2: {
Z extraParams[3];
for (auto i = start; i < stop; i++) {
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = xTadOffsets ? x + xTadOffsets[i] : x;
const auto yTad = yTadOffsets ? y + yTadOffsets[i] : y;
auto s = OpType::startingValue(xTad);
for (Nd4jLong i0 = 0; i0 < tadShape[0]; ++i0) {
for (Nd4jLong i1 = 0; i1 < tadShape[1]; ++i1) {
const auto xTadOffset = i0 * xTadStride[0] + i1 * xTadStride[1];
const auto yTadOffset = i0 * yTadStride[0] + i1 * yTadStride[1];
s = OpType::update(s, OpType::op(xTad[xTadOffset], yTad[yTadOffset], extraParams), extraParams);
}
}
z[i * zEws] = OpType::postProcess(s, tadLen, extraParams);
};
}
break;
//*********************************************//
case LoopKind::RANK3: {
Z extraParams[3];
for (auto i = start; i < stop; i++) {
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = xTadOffsets ? x + xTadOffsets[i] : x;
const auto yTad = yTadOffsets ? y + yTadOffsets[i] : y;
auto s = OpType::startingValue(xTad);
for (Nd4jLong i0 = 0; i0 < tadShape[0]; ++i0) {
for (Nd4jLong i1 = 0; i1 < tadShape[1]; ++i1) {
for (Nd4jLong i2 = 0; i2 < tadShape[2]; ++i2) {
const auto xTadOffset = i0 * xTadStride[0] + i1 * xTadStride[1] + i2 * xTadStride[2];
const auto yTadOffset = i0 * yTadStride[0] + i1 * yTadStride[1] + i2 * yTadStride[2];
s = OpType::update(s, OpType::op(xTad[xTadOffset], yTad[yTadOffset], extraParams), extraParams);
}
}
}
z[i * zEws] = OpType::postProcess(s, tadLen, extraParams);
};
}
break;
//*********************************************//
case LoopKind::RANK4: {
Z extraParams[3];
for (auto i = start; i < stop; i++) {
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = xTadOffsets ? x + xTadOffsets[i] : x;
const auto yTad = yTadOffsets ? y + yTadOffsets[i] : y;
auto s = OpType::startingValue(xTad);
for (Nd4jLong i0 = 0; i0 < tadShape[0]; ++i0) {
for (Nd4jLong i1 = 0; i1 < tadShape[1]; ++i1) {
for (Nd4jLong i2 = 0; i2 < tadShape[2]; ++i2) {
for (Nd4jLong i3 = 0; i3 < tadShape[3]; ++i3) {
const auto xTadOffset = i0 * xTadStride[0] + i1 * xTadStride[1] + i2 * xTadStride[2] + i3 * xTadStride[3];
const auto yTadOffset = i0 * yTadStride[0] + i1 * yTadStride[1] + i2 * yTadStride[2] + i3 * yTadStride[3];
s = OpType::update(s, OpType::op(xTad[xTadOffset], yTad[yTadOffset], extraParams), extraParams);
}
}
}
}
z[i * zEws] = OpType::postProcess(s, tadLen, extraParams);
};
}
break;
//*********************************************//
case LoopKind::RANK5: {
Z extraParams[3];
for (auto i = start; i < stop; i++) {
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = xTadOffsets ? x + xTadOffsets[i] : x;
const auto yTad = yTadOffsets ? y + yTadOffsets[i] : y;
auto s = OpType::startingValue(xTad);
for (Nd4jLong i0 = 0; i0 < tadShape[0]; ++i0) {
for (Nd4jLong i1 = 0; i1 < tadShape[1]; ++i1) {
for (Nd4jLong i2 = 0; i2 < tadShape[2]; ++i2) {
for (Nd4jLong i3 = 0; i3 < tadShape[3]; ++i3) {
for (Nd4jLong i4 = 0; i4 < tadShape[4]; ++i4) {
const auto xTadOffset = i0 * xTadStride[0] + i1 * xTadStride[1] + i2 * xTadStride[2] + i3 * xTadStride[3] + i4 * xTadStride[4];
const auto yTadOffset = i0 * yTadStride[0] + i1 * yTadStride[1] + i2 * yTadStride[2] + i3 * yTadStride[3] + i4 * yTadStride[4];
s = OpType::update(s, OpType::op(xTad[xTadOffset], yTad[yTadOffset], extraParams), extraParams);
}
}
}
}
}
z[i * zEws] = OpType::postProcess(s, tadLen, extraParams);
};
}
break;
//*********************************************//
default: {
uint castXTadShapeInfo[MAX_RANK];
const bool canCastXTad = sd::DataTypeUtils::castShapeInfo<uint>(xTadShapeInfo, castXTadShapeInfo);
if (shape::haveSameShapeAndStrides(xTadShapeInfo, yTadShapeInfo)) {
Z extraParams[3];
for (auto i = start; i < stop; i++) {
2019-06-06 14:21:15 +02:00
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = xTadOffsets ? x + xTadOffsets[i] : x;
const auto yTad = yTadOffsets ? y + yTadOffsets[i] : y;
auto s = OpType::startingValue(xTad);
2019-06-06 14:21:15 +02:00
for (Nd4jLong j = 0; j < tadLen; ++j) {
const auto tadOffset = shape::indexOffset(j, xTadShapeInfo, castXTadShapeInfo, canCastXTad);
s = OpType::update(s, OpType::op(xTad[tadOffset], yTad[tadOffset], extraParams), extraParams);
}
2019-06-06 14:21:15 +02:00
z[i * zEws] = OpType::postProcess(s, tadLen, extraParams);
};
2019-06-06 14:21:15 +02:00
}
else {
uint castYTadShapeInfo[MAX_RANK];
const bool canCastYTad = sd::DataTypeUtils::castShapeInfo<uint>(yTadShapeInfo, castYTadShapeInfo);
2019-06-06 14:21:15 +02:00
Z extraParams[3];
for (auto i = start; i < stop; i++) {
2019-06-06 14:21:15 +02:00
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = xTadOffsets ? x + xTadOffsets[i] : x;
const auto yTad = yTadOffsets ? y + yTadOffsets[i] : y;
auto s = OpType::startingValue(xTad);
2019-06-06 14:21:15 +02:00
for (Nd4jLong j = 0; j < tadLen; ++j) {
const auto xTadOffset = shape::indexOffset(j, xTadShapeInfo, castXTadShapeInfo, canCastXTad);
const auto yTadOffset = shape::indexOffset(j, yTadShapeInfo, castYTadShapeInfo, canCastYTad);
s = OpType::update(s, OpType::op(xTad[xTadOffset], yTad[yTadOffset], extraParams), extraParams);
}
z[i * zEws] = OpType::postProcess(s, tadLen, extraParams);
};
2019-06-06 14:21:15 +02:00
}
}
}
}
2019-06-06 14:21:15 +02:00
//////////////////////////////////////////////////////////////////////////////
template<typename X, typename Z>
template <typename OpType>
Legacy API changes (#441) * initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * another initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * one more initial commit Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next step Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored buffer() and shapeInfo() methods usage with NDArray class. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt Graph class methods to use const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt choose op to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt where op shape method to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt lstsq op to use constant empty shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt matrix_diag_part op shape routine to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt determinant ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt mean_pairwssqerr_loss ops to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for loss ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt log_loss op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt dilation2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted deconv2d ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted dynamicRNN op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for lstm layer ops. Signed-off-by: shugeo <sgazeos@gmail.com> * few updates Signed-off-by: raver119@gmail.com <raver119@gmail.com> * first cuda tweak Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Adopt constant shapes for sconv2d ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes for gru ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt constant shapes with shape methods for segment ops and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with unsorted_segment_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted constant shapes with gamma op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods of reduce_stddev ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape methods for reduce_* ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt shape method for squeeze op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt strided_slice shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored concat op shape method to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted shape method for mirror_pad op. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted split op shape method. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted tile ops shape methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Added const cast for mkldnn routines handles. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored logSoftMaxForVector_ routine to conform with proper data and shape pointer casts. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetic changes to proper usage of constant pointers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple shape comparators for strides and addBias helpers to proper use data pointers with inplace option. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored depthToSpace helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored histogram helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored im2col helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored gather and gatherND helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage on percentile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed gather shape with helpers and range buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with space to depth helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage and constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with LUP decomposition> Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored onehot_ helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pad and prefix to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactoed softmax helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed space to batch helpers to use buffers properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed stack and split helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with sparse to dense helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with mindistance_ helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with tile helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with legacy pairwise bool ops. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored a couple of methods to adopt constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed broadcasting with constant shape." Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const usage with inplace reverse and constant shapes with legacy reduction. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored sort to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected sort for constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed constant shape usage with special methods. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored Context to conform with constant shape usage. Signed-off-by: shugeo <sgazeos@gmail.com> * CUDA broadcasting headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * pairwise/indexreduce/random headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored native ops to adopt constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * legacy reduce3/scalar headers Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected pullRow signature and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected routines to proper use of constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored legacy ops tests to use constant shapes properly. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with NDArray tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed native ops tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed special concat routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with test. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed buffer usage with a test. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored TAD.h and tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored calcStrides* routines to use constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed miscelaneous errors with constant shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Corrected definitions for declared functions. Signed-off-by: shugeo <sgazeos@gmail.com> * NativeOps const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed const shapes with shape routines. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed shape method for broadcastable case. Signed-off-by: shugeo <sgazeos@gmail.com> * few more const changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * xw_plus_b BP shape fn restored Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed signatures with broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Repaired backprops shape methods for a set of operations. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored broadcast bool for cuda. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods for 3 args with const qualifier. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed a couple of kernel signatures for broadcasting. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels signatures for const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise methods to persistent buffers and shapes usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopt const to buffers and shapes with scalar kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored indexreduce kernels signatures to use const buffers and shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored pairwise bool kernels to adopt cons shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored random special ops to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored native ops to conform with const shapes and buffers under cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Cosmetical changes only. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes and buffers error. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected start pos routine. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored methods to conform with const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored helpers to use proper methods instead. Signed-off-by: shugeo <sgazeos@gmail.com> * bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * next bunch of changes Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed execScalar declaration. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected const shape cases with sort and so on. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes for sort. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored kernel declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernel declarations to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernels declarations to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed segment helpers kernels declarations and so on to adopt const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with segment and solve helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed kernel declaration with adjustWeight helper. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed cuda implementations for constant shape helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted const shape usage with kernels. Signed-off-by: shugeo <sgazeos@gmail.com> * Adopted top_k kernels to use const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected kernels declarations to adopt const shapes with helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored NDArray definitions to adopt const shapes and buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shapes with image suppression helpers. Signed-off-by: shugeo <sgazeos@gmail.com> * Slight improvement with buffers. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored buffer usage with tests. Signed-off-by: shugeo <sgazeos@gmail.com> * Fixed const shape usage with definitions. Signed-off-by: shugeo <sgazeos@gmail.com> * minor updates on cpu side Signed-off-by: raver119@gmail.com <raver119@gmail.com> * Refactored const shape usage with ConstantDescritor and native ops with cuda platform. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored tear and tile kernels to adopt with const shapes. Signed-off-by: shugeo <sgazeos@gmail.com> * softmax_loop fix Signed-off-by: raver119 <raver119@gmail.com> * update missing signature Signed-off-by: raver119@gmail.com <raver119@gmail.com> * softmax again Signed-off-by: raver119@gmail.com <raver119@gmail.com> * few more missing consts Signed-off-by: raver119 <raver119@gmail.com> * new methods updated Signed-off-by: raver119@gmail.com <raver119@gmail.com> Co-authored-by: shugeo <sgazeos@gmail.com>
2020-05-09 07:06:14 +02:00
void sd::Reduction3Loops<X, Z>::loopReduce3All(const X* x, const Nd4jLong* xShapeInfo,
const X* y, const Nd4jLong* yShapeInfo,
Z* z, const Nd4jLong* zShapeInfo,
const Nd4jLong* xTadShapeInfo, const Nd4jLong* xTadOffsets,
const Nd4jLong* yTadShapeInfo, const Nd4jLong* yTadOffsets,
Z* extraParameters,
int64_t start, int64_t stop) {
// both tads have same shape, however strides and ews may differ
Z param0(OpType::startingValue(x)), param1(OpType::startingValue(x)), param2(extraParameters ? extraParameters[0] : OpType::startingValue(x));
const LoopKind::Kind kindOfLoop = LoopKind::deduceKindOfLoopTadXYZ(xTadShapeInfo, yTadShapeInfo, zShapeInfo);
const auto xTadEws = shape::elementWiseStride(xTadShapeInfo);
const auto yTadEws = shape::elementWiseStride(yTadShapeInfo);
const auto zEws = shape::elementWiseStride(zShapeInfo);
const auto zLen = shape::length(zShapeInfo);
const auto tadLen = shape::length(xTadShapeInfo);
const auto numXTads = shape::length(xShapeInfo) / tadLen;
const auto numYTads = shape::length(yShapeInfo) / tadLen;
const auto tadShape = shape::shapeOf(xTadShapeInfo);
const auto xTadStride = shape::stride(xTadShapeInfo);
const auto yTadStride = shape::stride(yTadShapeInfo);
const auto startVal = OpType::startingValue(x);
int numThreads = OmpLaunchHelper::tadThreads(tadLen, numXTads * numYTads);
switch (kindOfLoop) {
//*********************************************//
case LoopKind::EWS1: {
Z extraParams[3];
for (Nd4jLong ix = 0; ix < numXTads; ix++) {
for (Nd4jLong iy = 0; iy < numYTads; iy++) {
2019-06-06 14:21:15 +02:00
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = x + xTadOffsets[ix];
const auto yTad = y + yTadOffsets[iy];
const auto zInd = ix * numYTads + iy;
auto s = startVal;
for (Nd4jLong j = 0; j < tadLen; ++j)
s = OpType::update(s, OpType::op(xTad[j], yTad[j], extraParams), extraParams);
z[zInd] = OpType::postProcess(s, tadLen, extraParams);
}
};
}
break;
//*********************************************//
case LoopKind::EWSNONZERO: {
Z extraParams[3];
for (Nd4jLong ix = 0; ix < numXTads; ix++) {
for (Nd4jLong iy = 0; iy < numYTads; iy++) {
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = x + xTadOffsets[ix];
const auto yTad = y + yTadOffsets[iy];
const auto zInd = ix * numYTads + iy;
auto s = startVal;
for (Nd4jLong j = 0; j < tadLen; ++j)
s = OpType::update(s, OpType::op(xTad[j * xTadEws], yTad[j * yTadEws], extraParams), extraParams);
z[zInd * zEws] = OpType::postProcess(s, tadLen, extraParams);
}
};
}
break;
//*********************************************//
case LoopKind::RANK1: {
Z extraParams[3];
for (Nd4jLong ix = 0; ix < numXTads; ix++) {
for (Nd4jLong iy = 0; iy < numYTads; iy++) {
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = x + xTadOffsets[ix];
const auto yTad = y + yTadOffsets[iy];
const auto zInd = ix * numYTads + iy;
auto s = startVal;
2019-06-06 14:21:15 +02:00
for (Nd4jLong i0 = 0; i0 < tadLen; ++i0) {
2019-06-06 14:21:15 +02:00
const auto xTadOffset = i0 * xTadStride[0];
const auto yTadOffset = i0 * yTadStride[0];
s = OpType::update(s, OpType::op(xTad[xTadOffset], yTad[yTadOffset], extraParams), extraParams);
2019-06-06 14:21:15 +02:00
}
z[zInd * zEws] = OpType::postProcess(s, tadLen, extraParams);
}
};
}
break;
//*********************************************//
case LoopKind::RANK2: {
Z extraParams[3];
for (Nd4jLong ix = 0; ix < numXTads; ix++) {
for (Nd4jLong iy = 0; iy < numYTads; iy++) {
2019-06-06 14:21:15 +02:00
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = x + xTadOffsets[ix];
const auto yTad = y + yTadOffsets[iy];
const auto zInd = ix * numYTads + iy;
auto s = startVal;
2019-06-06 14:21:15 +02:00
for (Nd4jLong i0 = 0; i0 < tadShape[0]; ++i0) {
for (Nd4jLong i1 = 0; i1 < tadShape[1]; ++i1) {
2019-06-06 14:21:15 +02:00
const auto xTadOffset = i0 * xTadStride[0] + i1 * xTadStride[1];
const auto yTadOffset = i0 * yTadStride[0] + i1 * yTadStride[1];
s = OpType::update(s, OpType::op(xTad[xTadOffset], yTad[yTadOffset], extraParams), extraParams);
2019-06-06 14:21:15 +02:00
}
}
z[zInd * zEws] = OpType::postProcess(s, tadLen, extraParams);
}
};
}
break;
2019-06-06 14:21:15 +02:00
//*********************************************//
case LoopKind::RANK3: {
Z extraParams[3];
for (Nd4jLong ix = 0; ix < numXTads; ix++) {
for (Nd4jLong iy = 0; iy < numYTads; iy++) {
2019-06-06 14:21:15 +02:00
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = x + xTadOffsets[ix];
const auto yTad = y + yTadOffsets[iy];
const auto zInd = ix * numYTads + iy;
auto s = startVal;
2019-06-06 14:21:15 +02:00
for (Nd4jLong i0 = 0; i0 < tadShape[0]; ++i0) {
for (Nd4jLong i1 = 0; i1 < tadShape[1]; ++i1) {
for (Nd4jLong i2 = 0; i2 < tadShape[2]; ++i2) {
2019-06-06 14:21:15 +02:00
const auto xTadOffset = i0 * xTadStride[0] + i1 * xTadStride[1] + i2 * xTadStride[2];
const auto yTadOffset = i0 * yTadStride[0] + i1 * yTadStride[1] + i2 * yTadStride[2];
s = OpType::update(s, OpType::op(xTad[xTadOffset], yTad[yTadOffset], extraParams), extraParams);
2019-06-06 14:21:15 +02:00
}
}
}
z[zInd * zEws] = OpType::postProcess(s, tadLen, extraParams);
}
};
}
break;
2019-06-06 14:21:15 +02:00
//*********************************************//
case LoopKind::RANK4: {
Z extraParams[3];
for (Nd4jLong ix = 0; ix < numXTads; ix++) {
for (Nd4jLong iy = 0; iy < numYTads; iy++) {
2019-06-06 14:21:15 +02:00
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = x + xTadOffsets[ix];
const auto yTad = y + yTadOffsets[iy];
const auto zInd = ix * numYTads + iy;
auto s = startVal;
2019-06-06 14:21:15 +02:00
for (Nd4jLong i0 = 0; i0 < tadShape[0]; ++i0) {
for (Nd4jLong i1 = 0; i1 < tadShape[1]; ++i1) {
for (Nd4jLong i2 = 0; i2 < tadShape[2]; ++i2) {
for (Nd4jLong i3 = 0; i3 < tadShape[3]; ++i3) {
2019-06-06 14:21:15 +02:00
const auto xTadOffset = i0 * xTadStride[0] + i1 * xTadStride[1] + i2 * xTadStride[2] + i3 * xTadStride[3];
const auto yTadOffset = i0 * yTadStride[0] + i1 * yTadStride[1] + i2 * yTadStride[2] + i3 * yTadStride[3];
s = OpType::update(s, OpType::op(xTad[xTadOffset], yTad[yTadOffset], extraParams), extraParams);
2019-06-06 14:21:15 +02:00
}
}
}
}
z[zInd * zEws] = OpType::postProcess(s, tadLen, extraParams);
}
};
}
break;
2019-06-06 14:21:15 +02:00
//*********************************************//
case LoopKind::RANK5: {
Z extraParams[3];
for (Nd4jLong ix = 0; ix < numXTads; ix++) {
for (Nd4jLong iy = 0; iy < numYTads; iy++) {
2019-06-06 14:21:15 +02:00
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = x + xTadOffsets[ix];
const auto yTad = y + yTadOffsets[iy];
const auto zInd = ix * numYTads + iy;
auto s = startVal;
2019-06-06 14:21:15 +02:00
for (Nd4jLong i0 = 0; i0 < tadShape[0]; ++i0) {
for (Nd4jLong i1 = 0; i1 < tadShape[1]; ++i1) {
for (Nd4jLong i2 = 0; i2 < tadShape[2]; ++i2) {
for (Nd4jLong i3 = 0; i3 < tadShape[3]; ++i3) {
for (Nd4jLong i4 = 0; i4 < tadShape[4]; ++i4) {
2019-06-06 14:21:15 +02:00
const auto xTadOffset = i0 * xTadStride[0] + i1 * xTadStride[1] + i2 * xTadStride[2] + i3 * xTadStride[3] + i4 * xTadStride[4];
const auto yTadOffset = i0 * yTadStride[0] + i1 * yTadStride[1] + i2 * yTadStride[2] + i3 * yTadStride[3] + i4 * yTadStride[4];
s = OpType::update(s, OpType::op(xTad[xTadOffset], yTad[yTadOffset], extraParams), extraParams);
2019-06-06 14:21:15 +02:00
}
}
}
}
}
z[zInd * zEws] = OpType::postProcess(start, tadLen, extraParams);
2019-06-06 14:21:15 +02:00
}
};
2019-06-06 14:21:15 +02:00
}
break;
2019-06-06 14:21:15 +02:00
//*********************************************//
default: {
uint castXTadShapeInfo[MAX_RANK];
const bool canCastXTad = sd::DataTypeUtils::castShapeInfo<uint>(xTadShapeInfo, castXTadShapeInfo);
2019-06-06 14:21:15 +02:00
if (shape::haveSameShapeAndStrides(xTadShapeInfo, yTadShapeInfo)) {
Z extraParams[3];
for (Nd4jLong ix = 0; ix < numXTads; ix++) {
for (Nd4jLong iy = 0; iy < numYTads; iy++) {
2019-06-06 14:21:15 +02:00
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = x + xTadOffsets[ix];
const auto yTad = y + yTadOffsets[iy];
const auto zInd = ix * numYTads + iy;
auto s = startVal;
2019-06-06 14:21:15 +02:00
for (Nd4jLong j = 0; j < tadLen; ++j) {
const auto tadOffset = shape::indexOffset(j, xTadShapeInfo, castXTadShapeInfo, canCastXTad);
s = OpType::update(s, OpType::op(xTad[tadOffset], yTad[tadOffset], extraParams), extraParams);
2019-06-06 14:21:15 +02:00
}
z[zInd * zEws] = OpType::postProcess(s, tadLen, extraParams);
2019-06-06 14:21:15 +02:00
}
};
2019-06-06 14:21:15 +02:00
}
else {
uint castYTadShapeInfo[MAX_RANK];
const bool canCastYTad = sd::DataTypeUtils::castShapeInfo<uint>(yTadShapeInfo, castYTadShapeInfo);
2019-06-06 14:21:15 +02:00
Z extraParams[3];
for (Nd4jLong ix = 0; ix < numXTads; ix++) {
for (Nd4jLong iy = 0; iy < numYTads; iy++) {
2019-06-06 14:21:15 +02:00
extraParams[0] = param0;
extraParams[1] = param1;
extraParams[2] = param2;
const auto xTad = x + xTadOffsets[ix];
const auto yTad = y + yTadOffsets[iy];
const auto zInd = ix * numYTads + iy;
auto s = startVal;
2019-06-06 14:21:15 +02:00
for (Nd4jLong j = 0; j < tadLen; ++j) {
const auto xTadOffset = shape::indexOffset(j, xTadShapeInfo, castXTadShapeInfo, canCastXTad);
const auto yTadOffset = shape::indexOffset(j, yTadShapeInfo, castYTadShapeInfo, canCastYTad);
s = OpType::update(s, OpType::op(xTad[xTadOffset], yTad[yTadOffset], extraParams), extraParams);
2019-06-06 14:21:15 +02:00
}
z[zInd * zEws] = OpType::postProcess(s, tadLen, extraParams);
2019-06-06 14:21:15 +02:00
}
};
2019-06-06 14:21:15 +02:00
}
}
2019-06-06 14:21:15 +02:00
}
}
}
#endif //LIBND4J_LOOPS_H