/* ******************************************************************************
*
*
* 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.
*
* See the NOTICE file distributed with this work for additional
* information regarding copyright ownership.
* 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
//
#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>
#include <helpers/ConstantTadHelper.h>
#include <system/openmp_pragmas.h>
#include <execution/Threads.h>
namespace sd {
template <typename X, typename Z, typename E>
class ND4J_EXPORT ReductionLoops {
protected:
public:
template <typename OpType>
static FORCEINLINE void loopReduce(sd::memory::Workspace* workspace, const X* x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int* dims, E* extraParams);
};
template <typename X, typename Z>
class ReductionFloatLoops : public ReductionLoops<X, Z, Z> {
public:
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);
template <typename OpType>
static void innerloopReduce(sd::memory::Workspace* workspace, const X* x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int* dims, Z* extraParams);
};
template <typename X, typename Z>
class ND4J_EXPORT ReductionBoolLoops : public ReductionLoops<X, Z, X> {
public:
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);
template <typename OpType>
static void innerloopReduce(sd::memory::Workspace* workspace, const X* x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int* dims, X* extraParams);
};
template <typename X, typename Z>
class ND4J_EXPORT ReductionLongLoops : public ReductionLoops<X, Z, X> {
public:
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);
template <typename OpType>
static void innerloopReduce(sd::memory::Workspace* workspace, const X* x, const Nd4jLong *xShapeInfo, Z* z, const Nd4jLong *zShapeInfo, const int* dims, X* extraParams);
};
template <typename X>
class ND4J_EXPORT ReductionSameLoops : public ReductionLoops<X, X, X> {
public:
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);
template <typename OpType>
static void innerloopReduce(sd::memory::Workspace* workspace, const X* x, const Nd4jLong *xShapeInfo, X* z, const Nd4jLong *zShapeInfo, const int* dims, X* extraParams);
};
template <typename X, typename Z>
class ND4J_EXPORT IndexReductionLoops {
private:
public:
static void wrapIndexReduce(int opNum, const void* x, const Nd4jLong* xShapeInfo, void* z, const Nd4jLong* zShapeInfo, const Nd4jLong* tadShapeInfo, const Nd4jLong* tadOffsets, void* extraParams);
template <typename OpType>
static void loopIndexReduce(const X* x, const Nd4jLong* xShapeInfo, Z* z, const Nd4jLong* zShapeInfo, const Nd4jLong* tadShapeInfo, const Nd4jLong* tadOffsets, X* extraParams);
};
template <typename X, typename Z, typename E>
class ND4J_EXPORT TransformLoops {
public:
template<typename OpType>
static FORCEINLINE void loopTransform(const X* x, const Nd4jLong* xShapeInfo, Z* z, const Nd4jLong* zShapeInfo, E* extraParams, uint64_t threadId, uint64_t numThreads);
};
template <typename X, typename Z>
class ND4J_EXPORT Reduction3Loops {
public:
template <typename OpType>
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);
template <typename OpType>
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);
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);
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);
template <typename OpType>
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);
template <typename OpType>
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);
};
//////////////////////////////////////////////////////////////////////////
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) {
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]);
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 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);
*z0 = OpType::postProcess(s, static_cast<Nd4jLong>(xAxis1), extraParams);
}
};
samediff::Threads::parallel_for(func, 0,xAxis0);
}
//////////////////////////////////////////////////////////////////////////
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) {
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 Nd4jLong tadLen = static_cast<Nd4jLong>(xAxis1 * xAxis2);
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 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);
*z0 = OpType::postProcess(s, tadLen, extraParams);
}
};
samediff::Threads::parallel_for(func, 0,xAxis0);
}
//////////////////////////////////////////////////////////////////////////
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) {
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]);
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);
}
}
};
samediff::Threads::parallel_for(func, 0,xAxis0,1, 0,xAxis1,1);
}
//////////////////////////////////////////////////////////////////////////
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) {
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 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);
}
};
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]);
const Nd4jLong tadLen = static_cast<Nd4jLong>(xAxis2 * xAxis3);
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 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);
}
}
};
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);
}
}
}
};
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);
}
};
samediff::Threads::parallel_for(func, 0,xAxis0);
}
//////////////////////////////////////////////////////////////////////////
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) {
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]);
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) {
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);
}
}
};
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]);
const Nd4jLong tadLen = static_cast<Nd4jLong>(xAxis3 * xAxis4);
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 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);
}
}
}
};
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);
}
}
}
}
};
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);
}
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);
}
//////////////////////////////////////////////////////////////////////////////
template <typename X, typename Z, typename E>
template <typename OpType>
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) {
const LoopKind::Kind kindOfLoop = LoopKind::deduceKindOfLoopXZ(xShapeInfo, zShapeInfo);
const Nd4jLong* xShape = shape::shapeOf(const_cast<Nd4jLong*>(xShapeInfo));
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;
switch (kindOfLoop) {
//*********************************************//
case LoopKind::EWS1: {
auto span = samediff::Span::build(threadId, numThreads, 0, len, 1);
int64_t start = span.startX(), stop = span.stopX();
for (auto i = start; i < stop; i++)
z[i] = OpType::op(x[i], extraParams);
}
break;
//*********************************************//
case LoopKind::EWSNONZERO: {
const uint xEws = shape::elementWiseStride(xShapeInfo);
const uint zEws = shape::elementWiseStride(zShapeInfo);
auto span = samediff::Span::build(threadId, numThreads, 0, len, 1);
int64_t start = span.startX(), stop = span.stopX();
for (auto i = start; i < stop; i++)
z[i * zEws] = OpType::op(x[i * xEws], extraParams);
}
break;
//*********************************************//
case LoopKind::Z_EWSNONZERO: {
const uint zEws = shape::elementWiseStride(zShapeInfo);
uint castXShapeInfo[MAX_RANK];
const bool canCastX = sd::DataTypeUtils::castShapeInfo<uint>(xShapeInfo, castXShapeInfo);
auto span = samediff::Span::build(threadId, numThreads, 0, len, 1);
int64_t start = span.startX(), stop = span.stopX();
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);
}
}
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;
//*********************************************//
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;
//*********************************************//
case LoopKind::RANK2: {
auto uXShape0 = static_cast<uint>(xShape[0]);
auto uXShape1 = static_cast<uint>(xShape[1]);
auto loop = samediff::ThreadsHelper::pickLoop2d(numThreads, uXShape0, uXShape1);
auto span = samediff::Span2::build(loop, threadId, numThreads, 0, uXShape0, 1, 0, uXShape1, 1);
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;
//*********************************************//
case LoopKind::RANK3: {
auto uXShape0 = xShape[0];
auto uXShape1 = xShape[1];
auto uXShape2 = xShape[2];
auto loop = samediff::ThreadsHelper::pickLoop2d(numThreads, uXShape0, uXShape1);
auto span = samediff::Span2::build(loop, threadId, numThreads, 0, uXShape0, 1, 0, uXShape1, 1);
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;
//*********************************************//
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;
//*********************************************//
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);
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];
for (Nd4jLong i3 = 0; i3 < uXShape3; ++i3) {
auto z1 = z0 + i3 * zStride[3];
auto x1 = x0 + i3 * xStride[3];
for (Nd4jLong i4 = 0; i4 < uXShape4; ++i4)
z[z1 + i4 * zStride[4]] = OpType::op(x[x1 + i4 * xStride[4]], extraParams);
}
}
}
break;
//*********************************************//
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);
}
}
}
}
//////////////////////////////////////////////////////////////////////////////
template<typename X, typename Z>
template <typename OpType>
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) {
// 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);
const Nd4jLong* xTadShapeInfo = nullptr, * yTadShapeInfo = nullptr, * xTadOffsets = nullptr, * yTadOffsets = nullptr;
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);
xTadShapeInfo = tadPackX.primaryShapeInfo();
yTadShapeInfo = tadPackY.primaryShapeInfo();
xTadOffsets = tadPackX.primaryOffsets();
yTadOffsets = tadPackY.primaryOffsets();
}
else if (yLen > xLen) {
tadPackY = sd::ConstantTadHelper::getInstance().tadForDimensions(yShapeInfo, dims, dimsLen);
xTadShapeInfo = xShapeInfo;
yTadShapeInfo = tadPackY.primaryShapeInfo();
yTadOffsets = tadPackY.primaryOffsets();
}
else {
tadPackX = sd::ConstantTadHelper::getInstance().tadForDimensions(xShapeInfo, dims, dimsLen);
yTadShapeInfo = yShapeInfo;
xTadShapeInfo = tadPackX.primaryShapeInfo();
xTadOffsets = tadPackX.primaryOffsets();
}
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 tadShape = shape::shapeOf(xTadShapeInfo);
const auto xTadStride = shape::stride(xTadShapeInfo);
const auto yTadStride = shape::stride(xTadShapeInfo);
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++) {
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) {
const auto tadOffset = shape::indexOffset(j, xTadShapeInfo, castXTadShapeInfo, canCastXTad);
s = OpType::update(s, OpType::op(xTad[tadOffset], yTad[tadOffset], extraParams), extraParams);
}
z[i * zEws] = OpType::postProcess(s, tadLen, extraParams);
};
}
else {
uint castYTadShapeInfo[MAX_RANK];
const bool canCastYTad = sd::DataTypeUtils::castShapeInfo<uint>(yTadShapeInfo, castYTadShapeInfo);
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) {
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);
};
}
}
}
}
//////////////////////////////////////////////////////////////////////////////
template<typename X, typename Z>
template <typename OpType>
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++) {
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;
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[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++) {
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 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[zInd * zEws] = OpType::postProcess(s, tadLen, extraParams);
}
};
}
break;
//*********************************************//
case LoopKind::RANK3: {
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 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[zInd * zEws] = OpType::postProcess(s, tadLen, extraParams);
}
};
}
break;
//*********************************************//
case LoopKind::RANK4: {
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 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[zInd * zEws] = OpType::postProcess(s, tadLen, extraParams);
}
};
}
break;
//*********************************************//
case LoopKind::RANK5: {
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 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[zInd * zEws] = OpType::postProcess(start, 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 (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) {
const auto tadOffset = shape::indexOffset(j, xTadShapeInfo, castXTadShapeInfo, canCastXTad);
s = OpType::update(s, OpType::op(xTad[tadOffset], yTad[tadOffset], extraParams), extraParams);
}
z[zInd * zEws] = OpType::postProcess(s, tadLen, extraParams);
}
};
}
else {
uint castYTadShapeInfo[MAX_RANK];
const bool canCastYTad = sd::DataTypeUtils::castShapeInfo<uint>(yTadShapeInfo, castYTadShapeInfo);
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) {
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[zInd * zEws] = OpType::postProcess(s, tadLen, extraParams);
}
};
}
}
}
}
}
#endif //LIBND4J_LOOPS_H