cavis/libnd4j/include/ops/declarable/helpers/cuda/lup.cu

1012 lines
49 KiB
Plaintext
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 raver119@gmail.com
//
#include <ops/declarable/helpers/top_k.h>
#include <helpers/MmulHelper.h>
#include <array/NDArrayFactory.h>
#include <graph/Status.h>
#include <helpers/ConstantTadHelper.h>
#include <helpers/ShapeUtils.h>
//#include <ops/declarable/generic/helpers/BroadcastHelper.h>
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
#include <cusolverDn.h>
#include <exceptions/cuda_exception.h>
2019-06-06 14:21:15 +02:00
namespace sd {
2019-06-06 14:21:15 +02:00
namespace ops {
namespace helpers {
// ------------------------------------------------------------------------------------------------------------------ //
// invert the second diagonal for lower diagonal matrix
template<typename T>
static __global__ void
invertKernelLow(void *invertedBuf, Nd4jLong *invertedShape, void *inputBuf, Nd4jLong *inputShape, Nd4jLong n) {
T* inverted = reinterpret_cast<T *>(invertedBuf);
T* input = reinterpret_cast<T*>(inputBuf);
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
auto start = threadIdx.x + blockIdx.x * blockDim.x;
auto step = blockDim.x * gridDim.x;
for (int i = start + 1; i < n; i += step) {
Nd4jLong pos[] = {i, i - 1};
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
Nd4jLong posX[] = {i, i};
Nd4jLong posY[] = {i - 1, i - 1};
auto xIndex = shape::getOffset(inputShape, pos);
auto dxIndex = shape::getOffset(inputShape, posX);
auto dyIndex = shape::getOffset(inputShape, posY);
auto zIndex = shape::getOffset(invertedShape, pos);
// invert lower triangular matrix
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
inverted[zIndex] = -input[xIndex] / (input[dxIndex] * input[dyIndex]);
// math::atomics::nd4j_atomicAdd(&inverted[zIndex], - input[xIndex] * inverted[iIndex] / input[dIndex]);
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
}
// ------------------------------------------------------------------------------------------------------------------ //
// invert diagonal vals to upper diagonal matrix
template<typename T>
static __global__ void
upvertKernel(void *invertedBuf, Nd4jLong *invertedShape, void *inputBuf, Nd4jLong *inputShape, Nd4jLong n) {
T *inverted = reinterpret_cast<T *>(invertedBuf);
T *input = reinterpret_cast<T *>(inputBuf);
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
auto start = threadIdx.x + blockIdx.x * blockDim.x;
auto step = blockDim.x * gridDim.x;
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
for (int i = start; i < n; i += step) {
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
Nd4jLong pos[] = {i, i};
auto xIndex = shape::getOffset(inputShape, pos);
auto zIndex = shape::getOffset(invertedShape, pos);
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
// math::atomics::nd4j_atomicDiv(&inverted[zIndex], input[xIndex]);
// invert diagonal elements
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
inverted[zIndex] /= input[xIndex];
}
}
2019-06-06 14:21:15 +02:00
// ------------------------------------------------------------------------------------------------------------------ //
// invert upper second diagonal
template<typename T>
static __global__ void
upvertKernelUp(void *invertedBuf, Nd4jLong *invertedShape, void *inputBuf, Nd4jLong *inputShape, Nd4jLong n) {
__shared__ T* inverted;
__shared__ T* input;
if (threadIdx.x == 0) {
inverted = reinterpret_cast<T *>(invertedBuf);
input = reinterpret_cast<T *>(inputBuf);
}
__syncthreads();
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
auto start = threadIdx.x + blockIdx.x * blockDim.x;
auto step = blockDim.x * gridDim.x;
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
for (int i = start; i < n - 1; i += step) {
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
Nd4jLong pos[] = {i, i + 1};
Nd4jLong posX[] = {i + 1, i + 1};
auto xIndex = shape::getOffset(inputShape, pos);
auto iIndex = shape::getOffset(invertedShape, posX);
auto zIndex = shape::getOffset(invertedShape, pos);
// invert upper matrix
math::atomics::nd4j_atomicAdd(&inverted[zIndex], -input[xIndex] * inverted[iIndex]); // / input[yIndex]);
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
//inputMatrix->t<T>(i, i + 1) * invertedMatrix->t<T>(i + 1, i + 1) / inputMatrix->t<T>(i, i)
}
2019-06-06 14:21:15 +02:00
}
// ------------------------------------------------------------------------------------------------------------------ //
template<typename T>
static __global__ void
invertLowKernel(void *invertedBuf, Nd4jLong *invertedShape, void *inputBuf, Nd4jLong *inputShape, Nd4jLong n) {
T *inverted = reinterpret_cast<T *>(invertedBuf);
T *input = reinterpret_cast<T *>(inputBuf);
if (threadIdx.x == 0) {
inverted = reinterpret_cast<T *>(invertedBuf);
input = reinterpret_cast<T *>(inputBuf);
}
__syncthreads();
auto tid = blockIdx.x * blockDim.x + threadIdx.x;
auto step = gridDim.x * blockDim.x;
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
for (int i = tid + 2; i < n; i += step) {
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
for (int j = i - 2; j >= 0; --j)
for (int k = 0; k < i; k++) {
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
Nd4jLong posZ[] = {i, j};
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
Nd4jLong posY[] = {k, j};
Nd4jLong posX[] = {i, k};
Nd4jLong posD[] = {i, i};
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
auto xIndex = shape::getOffset(inputShape, posX);
auto yIndex = shape::getOffset(invertedShape, posY);
auto dIndex = shape::getOffset(inputShape, posD);
auto zIndex = shape::getOffset(invertedShape, posZ);
// invert non-diagonal elements
math::atomics::nd4j_atomicAdd(&inverted[zIndex], -inverted[yIndex] * input[xIndex] / input[dIndex]);
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
}
}
// ------------------------------------------------------------------------------------------------------------------ //
// Invertion of upper triangular matrix non-diagonal elements when main and second diagonals already processed
template<typename T>
static __global__ void
invertUpKernel(void *invertedBuf, Nd4jLong *invertedShape, void *inputBuf, Nd4jLong *inputShape, Nd4jLong n) {
__shared__ T* inverted;
__shared__ T* input;
if (threadIdx.x == 0) {
inverted = reinterpret_cast<T *>(invertedBuf);;
input = reinterpret_cast<T *>(inputBuf);
}
__syncthreads();
auto tid = blockIdx.x * blockDim.x + threadIdx.x;
auto step = blockDim.x * gridDim.x;
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
for (int i = (int)n - tid - 2; i >= 0; i -= step) {
for (int j = i + 2; j < (int)n; j++)
for (int k = i; k < (int)n; k++) {
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
Nd4jLong posZ[] = {i, j};
Nd4jLong posY[] = {k, j};
Nd4jLong posX[] = {i, k};
// inversion with Joardan Gauss transformation
auto xIndex = shape::getOffset(inputShape, posX);
auto yIndex = shape::getOffset(invertedShape, posY);
auto zIndex = shape::getOffset(invertedShape, posZ);
// invert upper non-diagonal elements
math::atomics::nd4j_atomicAdd(&inverted[zIndex], -inverted[yIndex] * input[xIndex]);
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
}
}
// ------------------------------------------------------------------------------------------------------------------ //
// procedure to invert lower-triangular matrix.
// In current case lower triangular matrix has main diagonal with general values
//
template<typename T>
static void invertLowerMatrix_(LaunchContext *context, NDArray *inputMatrix, NDArray *invertedMatrix) {
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
int n = inputMatrix->rows();
invertedMatrix->setIdentity();
if (inputMatrix->isIdentityMatrix()) return;
auto stream = context->getCudaStream();
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
// invert lower matrix
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
// invert main diagonal
upvertKernel<T><<<1, n, 512, *stream>>>(invertedMatrix->specialBuffer(), invertedMatrix->specialShapeInfo(), inputMatrix->specialBuffer(), inputMatrix->specialShapeInfo(), n);
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
// invert the second diagonal
invertKernelLow<T><<<1, n, 512, *stream>>>(invertedMatrix->specialBuffer(), invertedMatrix->specialShapeInfo(), inputMatrix->specialBuffer(), inputMatrix->specialShapeInfo(), n);
// invert non-diagonal elements
invertLowKernel<T><<<n, n, 512, *stream>>>(invertedMatrix->specialBuffer(), invertedMatrix->specialShapeInfo(), inputMatrix->specialBuffer(), inputMatrix->specialShapeInfo(), n);
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
// ------------------------------------------------------------------------------------------------------------------ //
// caller for invert lower matrix routine
void invertLowerMatrix(LaunchContext *context, NDArray *inputMatrix, NDArray *invertedMatrix) {
NDArray::prepareSpecialUse({invertedMatrix}, {inputMatrix});
BUILD_SINGLE_SELECTOR(inputMatrix->dataType(), invertLowerMatrix_, (context, inputMatrix, invertedMatrix), FLOAT_NATIVE);
NDArray::registerSpecialUse({invertedMatrix}, {inputMatrix});
2019-06-06 14:21:15 +02:00
}
// ------------------------------------------------------------------------------------------------------------------ //
// procedure to invert upper-triangular matrix.
// In current case upper triangular matrix has main diagonal with all ones on it.
template<typename T>
static void invertUpperMatrix_(LaunchContext *context, NDArray* inputMatrix, NDArray* invertedMatrix) {
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
int n = inputMatrix->rows();
invertedMatrix->setIdentity();
auto stream = context->getCudaStream();
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
if (inputMatrix->isIdentityMatrix()) { // the inverse for I is I
return;
}
2019-06-06 14:21:15 +02:00
// invert upper matrix
// invert the second diagonal
upvertKernelUp<T><<<1, n, 512, *stream >>>(invertedMatrix->specialBuffer(), invertedMatrix->specialShapeInfo(),
inputMatrix->specialBuffer(), inputMatrix->specialShapeInfo(), n);
// invert other elements
invertUpKernel<T><<<n, n, 512, *stream >>>(invertedMatrix->specialBuffer(), invertedMatrix->specialShapeInfo(),inputMatrix->specialBuffer(), inputMatrix->specialShapeInfo(), n);
2019-06-06 14:21:15 +02:00
}
// ------------------------------------------------------------------------------------------------------------------ //
// invertion of upper triangular matrix - runner routine
void invertUpperMatrix(LaunchContext *context, NDArray *inputMatrix, NDArray *invertedMatrix) {
NDArray::prepareSpecialUse({invertedMatrix}, {inputMatrix});
BUILD_SINGLE_SELECTOR(invertedMatrix->dataType(), invertUpperMatrix_, (context, inputMatrix, invertedMatrix), FLOAT_NATIVE);
NDArray::prepareSpecialUse({invertedMatrix}, {inputMatrix});
2019-06-06 14:21:15 +02:00
}
// ------------------------------------------------------------------------------------------------------------------ //
// determinant kernel - accumulation product of all values on the main diagonal
template<typename T>
static __global__ void determinantKernel(T *compound, T *result, Nd4jLong len) {
auto start = blockIdx.x * blockDim.x + threadIdx.x;
auto step = blockDim.x * gridDim.x;
for (auto i = start; i < len; i += step) {
auto pos = i * len + i; //shape::getOffset(0, shape::shapeOf(shape), shape::stride(shape), di, 2);
// multiply all diagonal elements
math::atomics::nd4j_atomicMul(&result[0], compound[pos]);
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
}
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
// ------------------------------------------------------------------------------------------------------------------ //
// determinant logarithm - accumulation sum of all logarithm values on the main diagonal. All in logarithic values
// should be positive
template<typename T>
static __global__ void determinantLogKernel(T *compound, T *result, Nd4jLong len) {
auto start = blockIdx.x * blockDim.x + threadIdx.x;
auto step = blockDim.x * gridDim.x;
for (auto i = start; i < len; i += step) {
auto pos = i * len + i; //shape::getOffset(0, shape::shapeOf(shape), shape::stride(shape), di, 2);
// sum logs of all diagonal elements
math::atomics::nd4j_atomicAdd(result, math::nd4j_log<T,T>(math::nd4j_abs(compound[pos])));
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
}
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
// ------------------------------------------------------------------------------------------------------------------ //
// kernel to copy matrix with given shape to compound tensor with given pos
// output - a N-D tensor buffer with rank not less than 2, input - 2D square n x n matrix with n = rowLen
template<typename T, typename F>
static __global__ void
fillMatrix(void *output, Nd4jLong *outShape, void *input, Nd4jLong *inputShape, Nd4jLong pos, Nd4jLong rowLen) {
__shared__ F *matrix;
__shared__ T *inputBuf;
__shared__ Nd4jLong inputLen;
__shared__ Nd4jLong n2;
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
if (threadIdx.x == 0) {
matrix = reinterpret_cast<F*>(output);
inputBuf = reinterpret_cast<T*>(input);
inputLen = shape::length(inputShape);
n2 = rowLen * rowLen;
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
}
__syncthreads();
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
auto start = blockIdx.x * blockDim.x + threadIdx.x;
auto step = blockDim.x * gridDim.x;
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
for (int k = pos + start, j = start; j < n2; k += step, j += step) {
auto xIndex = shape::getIndexOffset(k, inputShape);
matrix[j] = (F) inputBuf[xIndex];
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
}
2019-06-06 14:21:15 +02:00
// ------------------------------------------------------------------------------------------------------------------ //
// same as above, but without type conversion
template<typename T>
static __global__ void
returnMatrix(void *output, Nd4jLong *outputShape, void *input, Nd4jLong *inputShape, Nd4jLong pos, Nd4jLong rowLen) {
__shared__ T* matrix;
__shared__ T* outputBuf;
__shared__ Nd4jLong outputLen;
__shared__ Nd4jLong n2;
if (threadIdx.x == 0) {
matrix = reinterpret_cast<T *>(input);
outputBuf = reinterpret_cast<T *>(output);
outputLen = shape::length(inputShape);
n2 = rowLen * rowLen;
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
__syncthreads();
auto start = blockIdx.x * blockDim.x + threadIdx.x;
auto step = blockDim.x * gridDim.x;
Merge master to upstream (#7945) * Shugeo strided slice zeros (#14) * Modified strided_slice op to properly work with empty-like shapes. * Fixed test for reduce_mean with empty-like input. * [WIP] Last merge (#15) * correct logsoftmax looss (#2) * Small SameDiff listener fix (#4) * Various fixes (#6) * #7839 Fix for asXMatrix and tests * #7866 EmbeddingSequenceLayer dtype fix + test * #7856 SameDiff save/load stream methods * #7859 RegressionEvaluation rank 4 fix + tests + axis configuration * EvaluationBinary 3d/4d * More evaluation 3d/4d tests * #7847 Evaluation empty checks * Small test ifx * #7848 Fix median edge case * Improve DL4J samediff layer tests * [WIP] FastText wrapper implemented (#8) * FastText implemented * Some fixes * Fix shapes for wordsNearest * Validation of input vectors * Fixes * Fixed test * Thread tagged * Some tweaks * setContextClassLoader for DeallocatorServiceThread * Numpy format tests (#1) * Various fixes (#11) * #7852 SameDiff gather fix * #7892 SameDiff placeholder to constant conversion * #7890 validate input rank for MLN/CG init methods * Fix broken permute shape calculation * Permute and gather fixes * Tests * #7850 LogSumExp fix + test * Handful of test fixes * Empty arrays with non-scalar shapes (#10) * minor rearrangements for lambdas * empty tensors with non-scalar shapes * numpy empty tensors with non-scalar shapes * few more empty tweaks * Small fixes * conv3d signature update * micro fix in batchnorm mkldnn * Import fixes * Fix * MKL-DNN update * Small fill fix * fill with empty input + test * Fixes * Small error improvement * Fix * one special test * couple of fixes for lstm * Rewrite TFGraphMapper.getNDArrayFromTensor to be maintainable and less error prone * Fixes * FP16 * Unsigned * BFloat16 * Fill op - empty tweaks * - couple of fixes for empty arrays construction - stack updated * strided slice fix * one transform test * provide method for reducing shapeInfo in case of input array is empty * Fixed reduceAlongDimensions to use empty input properly. * couple of broadcast tests * couple of tests broadcast tests + tweak to make them pass * add check of non-empty to methods producing sub-arrays * Fixed reshapeC with zeros in shape. * complete empty check in reduce_... legacy ops * Concat and cumsum/prod * Tweak to empty shape inference on import * add empty check to the rest of reduce legacy ops * one more test * correct typo in evalReduceShapeInfoEmpty * Added tests for reduce_* ops to tests with zero shapes. * few more tests for empty reductions * Fixed strided_slice op with empty case and tests. * one more empty reduction test * Fixed strided_slice test. * add empty check to NDArray::reshapei * infOrMax * empty min/max with infinity tests * made unstack working correctly with empty arrays * few IndexReduce tests + tweaks for empty shapes * add test for empty concat * few tests fixed * Validation fix for reductions on empty shapes * Reverse fix * Reduction shape calc fixes * SameDiff.generateOutputVariable: don't use shape function to determine number of outputs * Range fix * - NDArray constructor updated for scalars/empty arrays - few tests fixed * More fixes * Empty creator fixes * concat fix * concat fix * TF import tests: allow 'both all NaN' and 'both all inf' to pass * Slice, zero fraction, and reshape fixes * transpose, gather * Zero fraction * scalar cast fix * Empty reduction axis support * few more tests fixed * Fixed input checks conforming with TF for concat op and tests. * few tests fixed * matmul scalar shape fix * Fixed checkout for data type and scalarity with concat to allow non-empty scalars with vector concats. * broadcast bool fix * few more tests * few more tests * correct evalReduceShapeInfoEmpty * argmax/argmin + tests * one more empty edge case + one more test * argmax/argmin/realdiv_bp tweaks * empty reshape test + fix * Helper fixes * Small fixes * Gather test fix * Gather test fix * Small fixes * reduce scalar zero values * scalar mean workaround * Remove debug code * along dim mean workaround * one more test * - equalsTo() tweak for empty arrays - one more test * broadcast tweaks * [WIP] Fixing outstanding issues for NLP (#9) * Avoid using not-inited objects * Test fixed. * Redundant method avoided for models like FastText * KMeans++ implementation * KMeans++ implementation * Disable parallel execution * KMeans++ * Tests * Dev branch merge (#16) * SameDiff: convertDataType and gradient check util improvements (#12) * GradCheck util improvements * StopGradient constructor + test * SameDiff: Add datatype conversion * Javadoc and add DataType.isNumerical() * Small fix * Fix SameDiff TF import test cases intermediate naming (workaround for bad default) * TFGraphTestAllHelper: check intermediates in execution order * Add missing debug listener * [WIP] lstmBlock fix + other changes (#13) - fixes lstmBlock issue - changes NDArray method reshape(), permute(), transpose() by making them return instance instead of pointer - CheckNumerics op - fixes for ReduceBool IsInfOrNan & IsFinite * Small test fix * CheckNumerics op wrapper * Fix some issues on master (#17) * Fix DataVec test issue * Fix issue with dl4j SameDiff output layer * Dtype fix for lambda layers * #7912 BertIterator dtype fix (use float32 not global default) * [WIP] Next set of CUDA stuff (#7) New CUDA implementations and improvements * bad file * Dev branch master merge (#23) * SameDiff: convertDataType and gradient check util improvements (#12) * GradCheck util improvements * StopGradient constructor + test * SameDiff: Add datatype conversion * Javadoc and add DataType.isNumerical() * Small fix * Fix SameDiff TF import test cases intermediate naming (workaround for bad default) * TFGraphTestAllHelper: check intermediates in execution order * Add missing debug listener * [WIP] lstmBlock fix + other changes (#13) - fixes lstmBlock issue - changes NDArray method reshape(), permute(), transpose() by making them return instance instead of pointer - CheckNumerics op - fixes for ReduceBool IsInfOrNan & IsFinite * Small test fix * CheckNumerics op wrapper * Compatibility of deserialization (#18) Signed-off-by: Alexander Stoyakin <alexander.stoyakin@gmail.com> * SameDiff: add activation gradient checking support for debugging (#19) * SameDiff gradient checker: first pass on activation gradient checks * Fixes + tests for activation gradient checking * Javadoc * [WIP] Some nd4j data type corrections (#20) * Adjust data type * Set correct Data type. * Size of proper data type. * fix averaged cpu load (#22) * SameDiff ops, TF import and fixes (#24) * CheckNumerics tests + fixes + misc fixes Signed-off-by: AlexDBlack <blacka101@gmail.com> * Fake quant Signed-off-by: AlexDBlack <blacka101@gmail.com> * Fixes Signed-off-by: AlexDBlack <blacka101@gmail.com> * FakeQuantWithMinMaxArgs Signed-off-by: AlexDBlack <blacka101@gmail.com> * CheckNumerics fix Signed-off-by: AlexDBlack <blacka101@gmail.com> * Fix libnd4j ALL_INTS and ALL_FLOATS declaration (uint and bfloat types) Signed-off-by: AlexDBlack <blacka101@gmail.com> * Small fix Signed-off-by: AlexDBlack <blacka101@gmail.com> * Javadoc Signed-off-by: AlexDBlack <blacka101@gmail.com> * Exception tweak Signed-off-by: AlexDBlack <blacka101@gmail.com> * fix Signed-off-by: AlexDBlack <blacka101@gmail.com> * Fix for out of scope stack allocated var use Signed-off-by: AlexDBlack <blacka101@gmail.com> * Ignores Signed-off-by: AlexDBlack <blacka101@gmail.com> * Ignore for known failing test (already logged issue) Signed-off-by: AlexDBlack <blacka101@gmail.com> * Merge upstream to fork (#25) * Add thousand-separator commas to TotalParams (#7915) * Add thousand-separator commas to TotalParams The number of parameters can be quite large, and it would help the reading of the summary printout to have the TotalParams column & values at the bottom have thousand-separator-commas in them. * Add thousand-separator commas to MultiLayerNetwork Corresponding change to MultiLayerNetwork Signed-off-by: Jxtps Jxtps <jxtps435@gmail.com> * Update contributing and issue/PR templates (#7934) Signed-off-by: AlexDBlack <blacka101@gmail.com> * Fix link to AdaDelta paper (#7942) Fix link to AdaDelta paper hosted on matthewzeiler.com Signed-off-by: Jxtps * Fixes, and ignores for known/logged failing issues (#7943) Signed-off-by: AlexDBlack <blacka101@gmail.com> * SameDiff + DL4J/SameDiff: Multiple fixes (#28) * #7919 HDF5 attribute buffer length fix Signed-off-by: AlexDBlack <blacka101@gmail.com> * #7909 Arbiter constructor exception ux improvements Signed-off-by: AlexDBlack <blacka101@gmail.com> * #7925 RNN output layer length checks Signed-off-by: AlexDBlack <blacka101@gmail.com> * #7939 Add listener for validating inputs are not incorrectly modified Signed-off-by: AlexDBlack <blacka101@gmail.com> * #7939 Integrate NonInplaceValidationListener into tests * #7844 DL4J SameDiff fixes for variable minibatch size * DL4J SameDiff fixes - ensure gradient for input placeholder is available Signed-off-by: AlexDBlack <blacka101@gmail.com> * Tweaks to ExternalErrorsFunction - use placeholders, make more robust * Another fix * More fixes * More SameDiff/DL4J fixes * Scope out scalar array creation in BaseScalarOp * Remove debug code Signed-off-by: AlexDBlack <blacka101@gmail.com> * [WIP] Final dev branch merge (#29) * SameDiff: convertDataType and gradient check util improvements (#12) * GradCheck util improvements * StopGradient constructor + test * SameDiff: Add datatype conversion * Javadoc and add DataType.isNumerical() * Small fix * Fix SameDiff TF import test cases intermediate naming (workaround for bad default) * TFGraphTestAllHelper: check intermediates in execution order * Add missing debug listener * [WIP] lstmBlock fix + other changes (#13) - fixes lstmBlock issue - changes NDArray method reshape(), permute(), transpose() by making them return instance instead of pointer - CheckNumerics op - fixes for ReduceBool IsInfOrNan & IsFinite * Small test fix * CheckNumerics op wrapper * Compatibility of deserialization (#18) Signed-off-by: Alexander Stoyakin <alexander.stoyakin@gmail.com> * SameDiff: add activation gradient checking support for debugging (#19) * SameDiff gradient checker: first pass on activation gradient checks * Fixes + tests for activation gradient checking * Javadoc * [WIP] Some nd4j data type corrections (#20) * Adjust data type * Set correct Data type. * Size of proper data type. * fix averaged cpu load (#22) * [WIP] Multiple dataset iterators (#27) * Splitting dataset into arbitrary number * Fixes * Multiple split of iterator * Test * Test * Some fixes * signature change * one more tweak Signed-off-by: raver119 <raver119@gmail.com> * one more test for sequential use of DataSetIteratorSplitter Signed-off-by: raver119 <raver119@gmail.com> * Fixes * Fixes * one more test for Alexander Signed-off-by: raver119 <raver119@gmail.com> * Some fixes * Some fixes * one more test for Alexander Signed-off-by: raver119 <raver119@gmail.com> * minor test fix Signed-off-by: raver119 <raver119@gmail.com> * Some fixes * Some fixes * couple of assertions tweaked Signed-off-by: raver119 <raver119@gmail.com> * MDS splitter test :/ Signed-off-by: raver119 <raver119@gmail.com> * Minor refactoring * Multi dataset * Some fixes * More tests * Small number of test fixes/improvements (failures on CI) (#31) Signed-off-by: AlexDBlack <blacka101@gmail.com> * [WIP] More CUDA stuff (#26) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * LRN BP CUDA Signed-off-by: raver119 <raver119@gmail.com> * less memory Signed-off-by: raver119 <raver119@gmail.com> * Fixed bug with crop_and_resize op helper. * get rid of unnecessary index-calculation dunction Signed-off-by: Yurii <yurii@skymind.io> * Fixed sort with nth_element cuda-based helper. * Refactored nth_element. * Refactored nth_element op and tests. * Modified usage of dim array with sortTad routine. * Refactored main routine of helper for non_max_image_suppression op. * non_max_image_suppression op helper with cuda kernel implementation. Initial revision. * fix vol2col cuda kernel * meh Signed-off-by: raver119 <raver119@gmail.com> * topK concept Signed-off-by: raver119 <raver119@gmail.com> * unsorted topK with scanWitdh of 1 Signed-off-by: raver119 <raver119@gmail.com> * correct vol2col tests * sorted/unsorted topK Signed-off-by: raver119 <raver119@gmail.com> * implementation and fixing col2im/col2vol * Corrected usage flags with input/output with reverse op. * dup is const now Signed-off-by: raver119 <raver119@gmail.com> * percentile op Signed-off-by: raver119 <raver119@gmail.com> * group tests for mapool2d Signed-off-by: Yurii <yurii@skymind.io> * special test for george Signed-off-by: raver119 <raver119@gmail.com> * less threads for sortTad Signed-off-by: raver119 <raver119@gmail.com> * provide conv2d for cuda Signed-off-by: Yurii <yurii@skymind.io> * remove auther in sort tad kernel code Signed-off-by: Yurii <yurii@skymind.io> * provide depthwise_conv2d for cuda Signed-off-by: Yurii <yurii@skymind.io> * - max_pooling_with_argmax - null check for special use Signed-off-by: raver119 <raver119@gmail.com> * dts cuda Signed-off-by: raver119 <raver119@gmail.com> * provide sconv2d for cuda Signed-off-by: Yurii <yurii@skymind.io> * std cuda Signed-off-by: raver119 <raver119@gmail.com> * Refactored non_max_suppression op to conform TF implementation. * Improved suppression helper. * provide pooling3d for cuda Signed-off-by: Yurii <yurii@skymind.io> * minor lstm rearrangements Signed-off-by: raver119 <raver119@gmail.com> * more of minor lstm rearrangements Signed-off-by: raver119 <raver119@gmail.com> * (bi)dynamic_rnn Signed-off-by: raver119 <raver119@gmail.com> * templates init order Signed-off-by: raver119 <raver119@gmail.com> * Refactored non_max_suppression op. * Added cuda kernel for non_max_suppression. * CPU sort by key/value Signed-off-by: raver119 <raver119@gmail.com> * CPU sort TAD by key/value Signed-off-by: raver119 <raver119@gmail.com> * CPU sort TAD by key/value tests Signed-off-by: raver119 <raver119@gmail.com> * Eliminate compiler error with cuda implementation. * - repaired gradCheck in cuda - provide conv2d_bp for cuda Signed-off-by: Yurii <yurii@skymind.io> * missed signature Signed-off-by: raver119 <raver119@gmail.com> * provide depthwise_conv2d_bp for cuda Signed-off-by: Yurii <yurii@skymind.io> * Implementation of lup helper with cuda kernel. Initial commit. * further work on backprops for convolutions Signed-off-by: Yurii <yurii@skymind.io> * CUDA linear sort by key/val Signed-off-by: raver119 <raver119@gmail.com> * CUDA tad sort by key/val Signed-off-by: raver119 <raver119@gmail.com> * start providing of backprop for pooling2d/3d Signed-off-by: Yurii <yurii@skymind.io> * Added atomicAdd for bool datatype. * dynamic partition concept Signed-off-by: raver119 <raver119@gmail.com> * dynamic partition concept Signed-off-by: raver119 <raver119@gmail.com> * dynamic partition scalar CUDA Signed-off-by: raver119 <raver119@gmail.com> * important comment Signed-off-by: raver119 <raver119@gmail.com> * fix pooling2d/3d backprop helpers Signed-off-by: Yurii <yurii@skymind.io> * Added non-linear test with dynamic_partition. * Improved test for dynamic_partition. * dynamic_partition TAD concept Signed-off-by: raver119 <raver119@gmail.com> * - dynamic_partition TAD CUDA impl - dynamic_partition TAD CPU fix Signed-off-by: raver119 <raver119@gmail.com> * - rewrite cpu code for usampling2d/3d - write cuda code for usampling2d/3d Signed-off-by: Yurii <yurii@skymind.io> * dynamic_stitch CUDA vector case Signed-off-by: raver119 <raver119@gmail.com> * dynamic_stitch CUDA TAD case concept Signed-off-by: raver119 <raver119@gmail.com> * dynamic_stitch CUDA TAD case impl Signed-off-by: raver119 <raver119@gmail.com> * Added tests for dynamic_stitch 3D-4D cases. * minor tests tweaks Signed-off-by: raver119 <raver119@gmail.com> * Fixed type check for dynamic stitch. * min/max bp Signed-off-by: raver119 <raver119@gmail.com> * rewrite code for upsampling2d/3d cpu Signed-off-by: Yurii <yurii@skymind.io> * reduce min/max/norm_max bp Signed-off-by: raver119 <raver119@gmail.com> * lup implementation. Additional enhancements. * provide code for upsamling2d/3d backprop Signed-off-by: Yurii <yurii@skymind.io> * weightedCrossEntropyWithLogits Signed-off-by: raver119 <raver119@gmail.com> * Fixed template math atomicMul for 64bit ints. * Refactored dynamic_partition_bp op. * inverseBroadcast fix Signed-off-by: raver119 <raver119@gmail.com> * DynamicPartitionBP test datatype fixed. * - nd4j_atomicMul Windows fix - cpu/NDArrayLambda.hpp excluded from CUDA Signed-off-by: raver119 <raver119@gmail.com>
2019-06-27 17:37:04 +02:00
for (int k = pos + start, j = start; j < n2; k += step, j += step) {
auto zIndex = shape::getIndexOffset(k, outputShape);
outputBuf[zIndex] = matrix[j];
}
}
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
// ------------------------------------------------------------------------------------------------------------------ //
// fill up permutaion matrix kernel. Permutation matrix filled with zeros and ones
template<typename F>
static __global__ void fillUpPermutation(void *output, Nd4jLong *shape, int *source, int rowNum) {
F *permutation = reinterpret_cast<F *>(output);
auto start = blockIdx.x * blockDim.x + threadIdx.x;
auto step = blockDim.x * gridDim.x;
for (auto i = start; i < rowNum; i += step) {
int val = source[i] - 1;
Nd4jLong posF[] = {i, val};
auto pos = shape::getOffset(shape, posF);
permutation[pos] = F(1.f);
}
}
// ------------------------------------------------------------------------------------------------------------------ //
// LUP decomposition runner - using CUBLAS SOLVER
// if permutation is given, then using LUP decomposition, LU decomposition otherwise
// L - lower triangular, U - upper triangular, P - permutation matricies
// PA = LU
//
// input - A matrix nxn
// compound - C matrix L + U - I, or main diagonal and lower - L matrix, from the 2nd diagonal - U matrix
template<typename T, typename I>
static void lup_(LaunchContext *context, NDArray *input, NDArray *compound, NDArray *permutation) {
auto stream = context->getCudaStream();
auto n = input->rows();
std::lock_guard<std::mutex> lock(*LaunchContext::deviceMutex());
cusolverDnHandle_t* cusolverH = (cusolverDnHandle_t*)context->getCusolverHandle(); //nullptr;
// create solver handle
cusolverStatus_t status; //cusolverDnCreate(&cusolverH);
// if (CUSOLVER_STATUS_SUCCESS != status) {
// throw cuda_exception::build("Cannot create cuSolver handle", status);
// }
// set solver stream
status = cusolverDnSetStream(*cusolverH, *stream);
if (CUSOLVER_STATUS_SUCCESS != status) {
throw cuda_exception::build("Cannot set up stream for cuda solver", status);
}
int lwork = 0;
int *d_info = nullptr;
// allocate memory for permutation vector
auto err = cudaMalloc((void **) &d_info, sizeof(int));
if (err) {
throw cuda_exception::build("helpers::lup_: Cannot allocate memory for solver info buffer", err);
}
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
DataType dtype = input->dataType();
switch (dtype) { // there are two implementations with cublas for LUP decomposition - double and float
case DataType::DOUBLE: {
double *d_work = nullptr;
// compute internal buffer size
double *matrix = reinterpret_cast<double *>(input->specialBuffer());
status = cusolverDnDgetrf_bufferSize(
*cusolverH,
n,
n,
matrix,
n,
&lwork);
if (CUSOLVER_STATUS_SUCCESS != status) {
throw cuda_exception::build("helpers::lup_: Cannot create cuSolver handle", status);
}
err = cudaMalloc((void **) &d_work, sizeof(float) * lwork);
if (err) {
throw cuda_exception::build("helpers::lup_: Cannot allocate memory for solver data buffer",
err);
}
if (permutation == nullptr) {
status = cusolverDnDgetrf(
*cusolverH,
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
n,
n,
matrix,
n,
d_work,
nullptr,
d_info);
if (status != CUSOLVER_STATUS_SUCCESS) {
throw cuda_exception::build("helpers::lup_: LU factorization is failed due ",
status);
}
}
else {
NDArray permutVector('c', {n}, sd::DataType::INT32, context);
int* permutationBuf = permutVector.dataBuffer()->specialAsT<int>();
status = cusolverDnDgetrf(
*cusolverH,
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
n,
n,
matrix,
n,
d_work,
permutationBuf,
d_info);
if (status != CUSOLVER_STATUS_SUCCESS) {
throw cuda_exception::build("helpers::lup_: LU factorization is failed due ",
status);
}
if (permutation->rankOf() == 2) {
fillUpPermutation<double> <<< n, n, 1024, *stream >>>
(permutation->specialBuffer(), permutation->specialShapeInfo(), permutationBuf, n);
}
else {
permutVector.tickWriteDevice();
input->tickWriteDevice();
compound->assign(input);
permutation->assign(permutVector);
}
}
err = cudaFree(d_work);
if (err) {
throw cuda_exception::build("helpers::lup_: Cannot deallocate memory for solver data buffer",
err);
}
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
break;
case DataType::FLOAT32: {
float *matrix = reinterpret_cast<float*>(input->specialBuffer());
float *d_work = nullptr;
status = cusolverDnSgetrf_bufferSize(
*cusolverH,
n,
n,
matrix,
n,
&lwork);
if (CUSOLVER_STATUS_SUCCESS != status) {
throw cuda_exception::build("helpers::lup_: Cannot create cuSolver handle", status);
}
err = cudaMalloc((void **) &d_work, sizeof(float) * lwork);
if (err) {
throw cuda_exception::build("helpers::lup_: Cannot allocate memory for solver data buffer",
err);
}
if (permutation == nullptr)
status = cusolverDnSgetrf(
*cusolverH,
n,
n,
matrix,
n,
d_work,
nullptr,
d_info);
else {
NDArray permutVector('c', {n}, DataType::INT32, context);
int *permutationBuf = reinterpret_cast<int *>(permutVector.specialBuffer());
status = cusolverDnSgetrf(
*cusolverH,
n,
n,
matrix,
n,
d_work,
permutationBuf,
d_info);
if (permutation->rankOf() == 2) {
fillUpPermutation<I> <<< n, n, 128, *stream >>>
(permutation->specialBuffer(), permutation->specialShapeInfo(), permutationBuf, n);
permutation->tickWriteDevice();
}
else {
input->tickWriteDevice();
compound->assign(input);
permutation->assign(permutVector);
}
}
err = cudaFree(d_work);
if (err) {
throw cuda_exception::build("helpers::lup_: Cannot deallocate memory for solver data buffer",
err);
}
}
}
if (CUSOLVER_STATUS_SUCCESS != status) {
throw cuda_exception::build("helpers::lup_: Cannot make LU decomposition", status);
}
err = cudaFree(d_info);
if (err) {
throw cuda_exception::build("helpers::lup_: Cannot deallocate memory for solver info buffer", err);
}
// cusolverDnDestroy(cusolverH);
// NDArray::registerSpecialUse({input}, {input});
input->tickWriteDevice();
}
// ------------------------------------------------------------------------------------------------------------------ //
BUILD_DOUBLE_TEMPLATE(template void lup_,(LaunchContext * context, NDArray * input, NDArray * output, NDArray * permutation), FLOAT_NATIVE, INDEXING_TYPES);
template <typename T>
static __device__ void swapRows(T* matrix, Nd4jLong* shape, Nd4jLong theFirst, Nd4jLong theSecond, Nd4jLong n) {
if (theFirst != theSecond) {
for (auto i = 0; i < n; i++) {
Nd4jLong theFirstPos[] = {theFirst, i};
Nd4jLong theSecondPos[] = {theSecond, i};
auto theFirstIndex = shape::getOffset(shape, theFirstPos, 0);
auto theSecondIndex = shape::getOffset(shape, theSecondPos, 0);
math::nd4j_swap(matrix[theFirstIndex], matrix[theSecondIndex]);
}
}
}
template <typename T>
static __device__ void processColumns(Nd4jLong currentRow, Nd4jLong rowNum, T* compoundBuf, Nd4jLong* compoundShape) {
Nd4jLong xDiag[] = {currentRow, currentRow};
auto diagIndex = shape::getOffset(compoundShape, xDiag, 0);
for (auto j = currentRow + 1; j < rowNum; j++) {
Nd4jLong xRow[] = {j, currentRow};
auto rowIndex = shape::getOffset(compoundShape, xRow, 0);
compoundBuf[rowIndex] /= compoundBuf[diagIndex]; //output->t<T>(i, i);
for (auto k = currentRow + 1; k < rowNum; k++) {
Nd4jLong yRow[] = {j, k};
Nd4jLong yCol[] = {currentRow, k};
auto rowIndexY = shape::getOffset(compoundShape, yRow, 0);
auto colIndex = shape::getOffset(compoundShape, yCol, 0);
compoundBuf[rowIndexY] -= compoundBuf[rowIndex] * compoundBuf[colIndex];
}
}
}
template <typename T>
__device__ Nd4jLong argmaxCol(Nd4jLong column, T* compoundBuffer, Nd4jLong* compoundShape) {
auto rowNum = shape::sizeAt(compoundShape, 0);
Nd4jLong xInitial[] = {column, column};
auto xInitialIndex = shape::getOffset(compoundShape, xInitial, 0);
auto maxValue = T(0); //sd::math::nd4j_abs(compoundBuffer[xInitialIndex]);
auto result = -1LL;
for (auto rowCounter = column; rowCounter < rowNum; rowCounter++) {
Nd4jLong xPos[] = {rowCounter, column};
auto xIndex = shape::getOffset(compoundShape, xPos, 0);
if (sd::math::nd4j_abs(compoundBuffer[xIndex]) > maxValue) {
maxValue = sd::math::nd4j_max(maxValue, sd::math::nd4j_abs(compoundBuffer[xIndex]));
result = rowCounter;
}
}
return result;
}
template <typename T, typename I>
static __device__ int luNN(T* matrix, Nd4jLong* shape, I* permutation, Nd4jLong* permuShape, Nd4jLong n) {
for (auto i = 0; i < n - 1; i++) {
auto pivotIndex = argmaxCol(i, matrix, shape);
if (pivotIndex < 0) {
return -1;//throw std::runtime_error("helpers::luNN_: input matrix is singular.");
}
math::nd4j_swap(permutation[shape::getIndexOffset(i, permuShape)], permutation[shape::getIndexOffset(pivotIndex, permuShape)]);
swapRows(matrix, shape, (Nd4jLong)i, pivotIndex, n);
processColumns(i, n, matrix, shape);
}
return 0;
}
template <typename T, typename I>
static __global__ void luBatchedKernel(T* outputBuf, Nd4jLong* outputShape, I* permutations, Nd4jLong* permuShape,
Nd4jLong* outputTadShape, Nd4jLong* outputTadOffsets, Nd4jLong* permuTadShape, Nd4jLong* permuTadOffsets,
Nd4jLong batchNum) {
auto start = blockIdx.x * blockDim.x + threadIdx.x;
auto step = blockDim.x * gridDim.x;
for (auto b = start; b < batchNum; b += step) {
T* matrix = outputBuf + outputTadOffsets[b];
I* permutation = permutations + permuTadOffsets[b];
if (0 != luNN(matrix, outputTadShape, permutation, permuTadShape, shape::length(permuTadShape))) break;
}
}
template <typename T, typename I>
static void lu_(LaunchContext * context, NDArray* input, NDArray* output, NDArray* permutationVectors) {
auto n = input->sizeAt(-1);
auto stream = context->getCudaStream();
NDArray iota('c', {n}, permutationVectors->dataType(), context);// = NDArrayFactory::create(); // <int>('c', {n});
iota.linspace(0); iota.syncToDevice();
output->assign(input); // fill up output tensor with zeros
// output->tickWriteDevice();
permutationVectors->applyTrueBroadcast(sd::BroadcastOpsTuple::Assign(), iota, *permutationVectors, true, nullptr);
// permutationVectors->tickWriteDevice();
auto tads = ConstantTadHelper::getInstance()->tadForDimensions(output->shapeInfo(), {-2, -1});
auto permutaionTads = ConstantTadHelper::getInstance()->tadForDimensions(output->shapeInfo(), {-1});
auto batchNum = tads.numberOfTads();
luBatchedKernel<T,I><<<batchNum, 256, 1024, *stream>>>(reinterpret_cast<T*>(output->platformBuffer()),
output->specialShapeInfo(), reinterpret_cast<I*>(permutationVectors->platformBuffer()),
permutationVectors->specialShapeInfo(), tads.specialShapeInfo(), tads.specialOffsets(),
permutaionTads.specialShapeInfo(), permutaionTads.specialOffsets(), batchNum);
}
void lu(LaunchContext* context, NDArray* input, NDArray* output, NDArray* permutations) {
NDArray::prepareSpecialUse({output, permutations}, {input});
BUILD_DOUBLE_SELECTOR(input->dataType(), permutations->dataType(), lu_, (context, input, output, permutations), FLOAT_NATIVE, INDEXING_TYPES);
NDArray::registerSpecialUse({output, permutations}, {input});
}
// ------------------------------------------------------------------------------------------------------------------ //
template<typename T>
static int determinant_(sd::LaunchContext *context, NDArray *input, NDArray *output) {
Nd4jLong n = input->sizeAt(-1);
Nd4jLong n2 = n * n;
std::vector<int> dims();
auto packX = ConstantTadHelper::getInstance()->tadForDimensions(input->getShapeInfo(), {input->rankOf() - 2, input->rankOf() - 1});
//auto packZ = ConstantTadHelper::getInstance()->tadForDimensions(output->shapeInfo(), {output->rankOf() - 1});
// DataType dtype = input->dataType();
// if (dtype != DataType::DOUBLE)
// dtype = DataType::FLOAT32;
auto matrix = NDArrayFactory::create(input->ordering(), {n, n}, DataTypeUtils::fromT<T>(), context); //, block.getWorkspace());
auto det = NDArrayFactory::create<T>(1, context);
auto stream = context->getCudaStream();
NDArray::prepareSpecialUse({output}, {input});
dim3 launchDims(256, 256, 1024);
output->assign(1.f);
for (int e = 0; e < output->lengthOf(); e++) {
Nd4jLong pos = e * n2;
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
// if (matrix.dataType() == input->dataType())
fillMatrix<T, T><<<launchDims.x, launchDims.y, launchDims.z, *stream>>>(matrix.specialBuffer(), matrix.specialShapeInfo(), input->specialBuffer(), input->specialShapeInfo(), pos, n);
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
// else
// fillMatrix<T, float><<<launchDims.x, launchDims.y, launchDims.z, *stream>>>(matrix.specialBuffer(), matrix.specialShapeInfo(), input->specialBuffer(), input->specialShapeInfo(), pos, n);
lup_<T, int>(context, &matrix, nullptr, nullptr);
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
// else
// lup_<float>(context, &matrix, nullptr, nullptr);
auto offset = shape::getIndexOffset(e, output->shapeInfo());
auto inputBuf = reinterpret_cast<T *>(matrix.specialBuffer());
auto outputBuf = reinterpret_cast<T *>(output->specialBuffer()) + offset;
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
// if (matrix.dataType() == input->dataType())
determinantKernel<T> << < launchDims.x, launchDims.y, launchDims.z, *stream >> >
(inputBuf, outputBuf, n);
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
// else
// determinantKernel<T, float><<<launchDims.x, launchDims.y, launchDims.z, *stream >>> (inputBuf, outputBuf, n);
}
NDArray::registerSpecialUse({output}, {input});
return Status::OK();
}
int determinant(sd::LaunchContext *context, NDArray *input, NDArray *output) {
NDArray::prepareSpecialUse({output}, {input});
BUILD_SINGLE_SELECTOR(input->dataType(), return determinant_, (context, input, output), FLOAT_NATIVE);
NDArray::registerSpecialUse({output}, {input});
}
template<typename T>
int logAbsDeterminant_(LaunchContext *context, NDArray *input, NDArray *output) {
Nd4jLong n = input->sizeAt(-1);
Nd4jLong n2 = n * n;
std::vector<int> dims();
auto packX = ConstantTadHelper::getInstance()->tadForDimensions(input->getShapeInfo(), {input->rankOf() - 2, input->rankOf() - 1});
//auto packZ = ConstantTadHelper::getInstance()->tadForDimensions(output->shapeInfo(), {output->rankOf() - 1});
DataType dtype = input->dataType();
if (dtype != DataType::DOUBLE)
dtype = DataType::FLOAT32;
auto matrix = NDArrayFactory::create(input->ordering(), {n, n}, dtype, context); //, block.getWorkspace());
auto det = NDArrayFactory::create<T>(1, context);
auto stream = context->getCudaStream();
NDArray::prepareSpecialUse({output}, {input});
dim3 launchDims(256, 256, 1024);
output->assign(0.f);
for (int e = 0; e < output->lengthOf(); e++) {
Nd4jLong pos = e * n2;
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
// if (matrix.dataType() == input->dataType())
fillMatrix<T, T><<<launchDims.x, launchDims.y, launchDims.z, *stream>>>(matrix.specialBuffer(), matrix.specialShapeInfo(), input->specialBuffer(), input->specialShapeInfo(), pos, n);
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
// else
// fillMatrix<T, float><<<launchDims.x, launchDims.y, launchDims.z, *stream>>>(matrix.specialBuffer(), matrix.specialShapeInfo(), input->specialBuffer(), input->specialShapeInfo(), pos, n);
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
// if (matrix.dataType() == input->dataType())
lup_<T, int>(context, &matrix, nullptr, nullptr);
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
// else
// lup_<float>(context, &matrix, nullptr, nullptr);
auto offset = shape::getIndexOffset(e, output->shapeInfo());
auto inputBuf = reinterpret_cast<T *>(matrix.specialBuffer());
auto outputBuf = reinterpret_cast<T *>(output->specialBuffer()) + offset;
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
// if (matrix.dataType() == input->dataType())
determinantLogKernel<T><<<launchDims.x, launchDims.y, launchDims.z, *stream>>>(inputBuf, outputBuf, n);
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
// else
// determinantLogKernel<T, float><<<launchDims.x, launchDims.y, launchDims.z, *stream >>> (inputBuf, outputBuf, n);
}
NDArray::registerSpecialUse({output}, {input});
return Status::OK();
return ND4J_STATUS_OK;
}
int logAbsDeterminant(sd::LaunchContext *context, NDArray *input, NDArray *output) {
NDArray::prepareSpecialUse({output}, {input});
BUILD_SINGLE_SELECTOR(input->dataType(), return logAbsDeterminant_, (context, input, output), FLOAT_NATIVE);
NDArray::registerSpecialUse({output}, {input});
}
template<typename T>
static __global__ void
fillLowerUpperKernel(void *lowerBuf, Nd4jLong *lowerShape, void *upperBuf, Nd4jLong *upperShape,
void *matrixBuf, Nd4jLong *matrixShape, Nd4jLong n) {
__shared__ T *lowerMatrix;
__shared__ T *upperMatrix;
__shared__ T *matrix;
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
if (threadIdx.x == 0) {
lowerMatrix = reinterpret_cast<T *>(lowerBuf);
upperMatrix = reinterpret_cast<T *>(upperBuf);
matrix = reinterpret_cast<T *>(matrixBuf);
}
__syncthreads();
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
for (int k = blockIdx.x; k < n; k += gridDim.x) { // and then put all values under main diagonal on to it
for (int j = threadIdx.x; j < n; j += blockDim.x) {
Nd4jLong posX[] = {k, j};
Nd4jLong posD[] = {j, j};
auto xPos = shape::getOffset(lowerShape, posX);
auto yPos = shape::getOffset(upperShape, posX);
auto iPos = shape::getOffset(matrixShape, posX);
auto dPos = shape::getOffset(matrixShape, posD);
if (k >= j)
lowerMatrix[xPos] = matrix[iPos];//(k, j);
else
upperMatrix[yPos] = matrix[iPos]; //k, j);
}
}
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
template<typename T>
static int inverse_(sd::LaunchContext *context, NDArray *input, NDArray *output) {
auto n = input->sizeAt(-1);
auto n2 = n * n;
auto dtype = DataTypeUtils::fromT<T>(); //input->dataType();
// if (dtype != DataType::DOUBLE)
// dtype = DataType::FLOAT32;
NDArray matrix = NDArrayFactory::create('c', {n, n}, dtype, context);
NDArray upper = NDArrayFactory::create('c', {n, n}, dtype, context);
NDArray lower = NDArrayFactory::create('c', {n, n}, dtype, context);
NDArray compound = NDArrayFactory::create('c', {n, n}, dtype, context);
NDArray permutation = NDArrayFactory::create('c', {n, n}, dtype, context);
auto packX = sd::ConstantTadHelper::getInstance()->tadForDimensions(input->getShapeInfo(),
{input->rankOf() - 2,
input->rankOf() - 1});
auto packZ = sd::ConstantTadHelper::getInstance()->tadForDimensions(output->getShapeInfo(),
{output->rankOf() - 2,
output->rankOf() - 1});
auto stream = context->getCudaStream();
for (auto i = 0LL; i < packX.numberOfTads(); i++) {
fillMatrix<T, T><<<1, n2, 1024, *stream>>>(matrix.specialBuffer(), matrix.specialShapeInfo(), input->specialBuffer(), input->specialShapeInfo(), i * n2, n);
matrix.tickWriteDevice();
//compound.assign(matrix);
// if (matrix.dataType() == input->dataType())
lup_<T, int>(context, &matrix, nullptr, nullptr);
fillLowerUpperKernel<T><<<n, n, 1024, *stream>>>(lower.specialBuffer(), lower.specialShapeInfo(), upper.specialBuffer(), upper.specialShapeInfo(), matrix.specialBuffer(), matrix.specialShapeInfo(), n);
lower.tickWriteDevice();
upper.tickWriteDevice();
// lower.printIndexedBuffer("LOWER");
// upper.printIndexedBuffer("UPPER");
matrix.assign(0);
invertUpperMatrix(context, &upper, &matrix); // U^{-1}
matrix.tickWriteDevice();
// matrix.printIndexedBuffer("Upper Inverted");
compound.assign(0);
invertLowerMatrix(context, &lower, &compound); // L{-1}
compound.tickWriteDevice();
// compound.printIndexedBuffer("Lower Inverted");
// matrix.tickWriteDevice();
// compound.tickWriteDevice();
sd::MmulHelper::mmul(&matrix, &compound, &upper, 1.0, 0.0);
upper.tickWriteDevice();
// upper.printIndexedBuffer("Full inverted");
returnMatrix<T> <<<1, n2, 1024, *stream>>>(output->specialBuffer(), output->specialShapeInfo(), upper.specialBuffer(), upper.specialShapeInfo(), i * n2, n);
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
return Status::OK();
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
2019-06-06 14:21:15 +02:00
int inverse(sd::LaunchContext *context, NDArray *input, NDArray *output) {
NDArray::prepareSpecialUse({output}, {input});
BUILD_SINGLE_SELECTOR(input->dataType(), return inverse_, (context, input, output), FLOAT_NATIVE);
NDArray::registerSpecialUse({output}, {input});
[WIP] More of CUDA operations (#69) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * - gruCell_bp further Signed-off-by: Yurii <yurii@skymind.io> * - further work on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Inverse matrix cublas implementation. Partial working revision. * Separation of segment ops helpers. Max separation. * Separated segment_min ops. * Separation of segment_mean/sum/prod/sqrtN ops heleprs. * Fixed diagonal processing with LUP decomposition. * Modified inversion approach using current state of LU decomposition. * Implementation of matrix_inverse op with cuda kernels. Working revision. * Implemented sequence_mask cuda helper. Eliminated waste printf with matrix_inverse implementation. Added proper tests. * - further work on gruCell_bp (ff/cuda) Signed-off-by: Yurii <yurii@skymind.io> * comment one test for gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda static_rnn Signed-off-by: Yurii <yurii@skymind.io> * Refactored random_shuffle op to use new random generator. * Refactored random_shuffle op helper. * Fixed debug tests with random ops tests. * Implement random_shuffle op cuda kernel helper and tests. * - provide cuda scatter_update Signed-off-by: Yurii <yurii@skymind.io> * Implementation of random_shuffle for linear case with cuda kernels and tests. * Implemented random_shuffle with cuda kernels. Final revision. * - finally gruCell_bp is completed Signed-off-by: Yurii <yurii@skymind.io> * Dropout op cuda helper implementation. * Implemented dropout_bp cuda helper. * Implemented alpha_dropout_bp with cuda kernel helpers. * Refactored helper. * Implementation of suppresion helper with cuda kernels. * - provide cpu code fot hsvToRgb, rgbToHsv, adjustHue Signed-off-by: Yurii <yurii@skymind.io> * Using sort by value method. * Implementation of image.non_max_suppression op cuda-based helper. * - correcting and testing adjust_hue, adjust_saturation cpu/cuda code Signed-off-by: Yurii <yurii@skymind.io> * Added cuda device prefixes to declarations. * Implementation of hashcode op with cuda helper. Initital revision. * rnn cu impl removed Signed-off-by: raver119 <raver119@gmail.com>
2019-07-20 07:58:44 +02:00
}
2019-06-06 14:21:15 +02:00
bool checkCholeskyInput(sd::LaunchContext *context, NDArray const *input) {
return true;
}
2019-06-06 14:21:15 +02:00
template<typename F>
__global__ void fillBatchKernel(F **dArrayBatch, F *buf, Nd4jLong *offsets, Nd4jLong batchSize) {
auto start = blockIdx.x * blockDim.x + threadIdx.x;
auto step = blockDim.x * gridDim.x;
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
for (auto i = start; i < batchSize; i += step) {
dArrayBatch[i] = buf + offsets[i];
}
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
template<typename F>
__global__ void
adjustResultsKernel(F *dArray, Nd4jLong *shape, Nd4jLong *offsets, Nd4jLong batchSize, Nd4jLong n) {
//auto i = blockIdx.x * blockDim.x + threadIdx.x;
Nd4jLong *shapeOf = shape::shapeOf(shape);
Nd4jLong *strideOf = shape::stride(shape);
for (auto i = blockIdx.x; i < batchSize; i += gridDim.x) {
auto current = dArray + offsets[i];
for (auto r = threadIdx.x; r < n; r += blockDim.x) {
for (auto c = r + 1; c < n; c++) {
Nd4jLong posRC[] = {r, c};
auto pos = r * n + c; //shape::getOffset(0, shapeOf, strideOf, posRC, 2);
current[pos] = 0.;
}
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
}
}
template<typename F>
int cholesky__(LaunchContext *context, NDArray *input, NDArray *output, bool inplace) {
if (!inplace)
output->assign(input);
Shyrma temp (#131) * - specifying template instantiation for certain types in float16 and bloat16 Signed-off-by: Yurii <iuriish@yahoo.com> * - polishing bfloat16 and float16 member functions template specialization Signed-off-by: Yurii <iuriish@yahoo.com> * - rewrite and overload array +-*/ scalar and scalar +-*/ arr in NDAray class Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections which have to do with and rvalue lvalue conversions Signed-off-by: Yurii <iuriish@yahoo.com> * - provide move semantic in NDArray operators array +-/* array Signed-off-by: Yurii <iuriish@yahoo.com> * float16/bfloat16 tweaks Signed-off-by: raver119 <raver119@gmail.com> * one more tweak Signed-off-by: raver119 <raver119@gmail.com> * - make float16 and bfloat16 to compile successfully on cuda Signed-off-by: Yurii <iuriish@yahoo.com> * - do not use resources of view-like arrays when move semantics is applied Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of pointers in signatures NDArray methods 1 Signed-off-by: Yurii <iuriish@yahoo.com> * - correction of signature of NDArray::dup method Signed-off-by: Yurii <iuriish@yahoo.com> * - correction of signature of NDArray::reduceAlongDimension method Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyIndexReduce and applyTrueBroadcast methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyReduce3 and varianceAlongDimension methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::tensorsAlongDimension and diagonal methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::allTensorsAlongDimension Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::reduceAlongDimension 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyTransform 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyPairwiseTransform 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyBroadcast 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyTrueBroadcast 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyScalar and applyScalarArr Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::lambda methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::reduce3 methods 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of following NDArray methods: add/sub/mul/div row/column and fillAsTriangular Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::tileToShape methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::isShapeSameStrict method Signed-off-by: Yurii <iuriish@yahoo.com> * minor corrections in tests Signed-off-by: Yurii <iuriish@yahoo.com> * - replace reduce op in batchnorm mkldnn Signed-off-by: Yurii <iuriish@yahoo.com> * - add explicit templates instantiations for operator+(NDArray&&. const scalar) Signed-off-by: Yurii <iuriish@yahoo.com> * - corrections of casts in float16/bfloat16 Signed-off-by: Yurii <iuriish@yahoo.com> * - provide move semantics in following NDArray methods: transform, applyTrueBroadcast, transpose, reshape, permute Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of input array A duplicate in svd cuda op Signed-off-by: Yurii <iuriish@yahoo.com> * - avoid available bug in svd cuda API Signed-off-by: Yurii <iuriish@yahoo.com> * - add temporary global memory buffer in svd cuda when calcUV = false and m != n Signed-off-by: Yurii <iuriish@yahoo.com> * - remove test with blfoat16 type for betainC Signed-off-by: Yurii <iuriish@yahoo.com> * - resolve conflicts after master has been merged in Signed-off-by: Yurii <iuriish@yahoo.com> * - changed type of affected input array in fused_batch_norm Signed-off-by: Yurii <iuriish@yahoo.com> * - add several explicit type castings Signed-off-by: Yurii <iuriish@yahoo.com> * - add ND4J_EXPORT to operators Signed-off-by: Yurii <iuriish@yahoo.com> * - add explicit template types in instantiations of template arithm operators of NDArray class Signed-off-by: Yurii <iuriish@yahoo.com> * - one more test fix Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com>
2019-12-20 20:35:39 +01:00
auto tempOutput =output->dup();
cusolverDnHandle_t handle = nullptr;
auto n = input->sizeAt(-1);
auto n2 = n * n;
NDArray::prepareSpecialUse({output}, {input});
auto status = cusolverDnCreate(&handle);
if (CUSOLVER_STATUS_SUCCESS != status) {
throw cuda_exception::build("helpers::cholesky_: Cannot create solver handle", status);
}
F **dArrayBatch = nullptr;
auto packX = sd::ConstantTadHelper::getInstance()->tadForDimensions(tempOutput.getShapeInfo(),
Shyrma temp (#131) * - specifying template instantiation for certain types in float16 and bloat16 Signed-off-by: Yurii <iuriish@yahoo.com> * - polishing bfloat16 and float16 member functions template specialization Signed-off-by: Yurii <iuriish@yahoo.com> * - rewrite and overload array +-*/ scalar and scalar +-*/ arr in NDAray class Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections which have to do with and rvalue lvalue conversions Signed-off-by: Yurii <iuriish@yahoo.com> * - provide move semantic in NDArray operators array +-/* array Signed-off-by: Yurii <iuriish@yahoo.com> * float16/bfloat16 tweaks Signed-off-by: raver119 <raver119@gmail.com> * one more tweak Signed-off-by: raver119 <raver119@gmail.com> * - make float16 and bfloat16 to compile successfully on cuda Signed-off-by: Yurii <iuriish@yahoo.com> * - do not use resources of view-like arrays when move semantics is applied Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of pointers in signatures NDArray methods 1 Signed-off-by: Yurii <iuriish@yahoo.com> * - correction of signature of NDArray::dup method Signed-off-by: Yurii <iuriish@yahoo.com> * - correction of signature of NDArray::reduceAlongDimension method Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyIndexReduce and applyTrueBroadcast methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyReduce3 and varianceAlongDimension methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::tensorsAlongDimension and diagonal methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::allTensorsAlongDimension Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::reduceAlongDimension 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyTransform 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyPairwiseTransform 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyBroadcast 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyTrueBroadcast 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyScalar and applyScalarArr Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::lambda methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::reduce3 methods 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of following NDArray methods: add/sub/mul/div row/column and fillAsTriangular Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::tileToShape methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::isShapeSameStrict method Signed-off-by: Yurii <iuriish@yahoo.com> * minor corrections in tests Signed-off-by: Yurii <iuriish@yahoo.com> * - replace reduce op in batchnorm mkldnn Signed-off-by: Yurii <iuriish@yahoo.com> * - add explicit templates instantiations for operator+(NDArray&&. const scalar) Signed-off-by: Yurii <iuriish@yahoo.com> * - corrections of casts in float16/bfloat16 Signed-off-by: Yurii <iuriish@yahoo.com> * - provide move semantics in following NDArray methods: transform, applyTrueBroadcast, transpose, reshape, permute Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of input array A duplicate in svd cuda op Signed-off-by: Yurii <iuriish@yahoo.com> * - avoid available bug in svd cuda API Signed-off-by: Yurii <iuriish@yahoo.com> * - add temporary global memory buffer in svd cuda when calcUV = false and m != n Signed-off-by: Yurii <iuriish@yahoo.com> * - remove test with blfoat16 type for betainC Signed-off-by: Yurii <iuriish@yahoo.com> * - resolve conflicts after master has been merged in Signed-off-by: Yurii <iuriish@yahoo.com> * - changed type of affected input array in fused_batch_norm Signed-off-by: Yurii <iuriish@yahoo.com> * - add several explicit type castings Signed-off-by: Yurii <iuriish@yahoo.com> * - add ND4J_EXPORT to operators Signed-off-by: Yurii <iuriish@yahoo.com> * - add explicit template types in instantiations of template arithm operators of NDArray class Signed-off-by: Yurii <iuriish@yahoo.com> * - one more test fix Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com>
2019-12-20 20:35:39 +01:00
{tempOutput.rankOf() - 2,
tempOutput.rankOf() - 1});
const Nd4jLong batchSize = packX.numberOfTads();
int *dInfoArray = nullptr;
auto err = cudaMalloc((void **) &dArrayBatch, sizeof(F *) * batchSize);
if (err) {
throw cuda_exception::build("helpers::cholesky_: Cannot allocate memory for solver batch data buffer",
err);
}
err = cudaMalloc((void **) &dInfoArray, sizeof(int) * batchSize);
if (err) {
throw cuda_exception::build("helpers::cholesky_: Cannot allocate memory for solver errors buffer", err);
}
auto stream = context->getCudaStream();
fillBatchKernel<F> << < 1, batchSize, 128, *stream >> >
Shyrma temp (#131) * - specifying template instantiation for certain types in float16 and bloat16 Signed-off-by: Yurii <iuriish@yahoo.com> * - polishing bfloat16 and float16 member functions template specialization Signed-off-by: Yurii <iuriish@yahoo.com> * - rewrite and overload array +-*/ scalar and scalar +-*/ arr in NDAray class Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections which have to do with and rvalue lvalue conversions Signed-off-by: Yurii <iuriish@yahoo.com> * - provide move semantic in NDArray operators array +-/* array Signed-off-by: Yurii <iuriish@yahoo.com> * float16/bfloat16 tweaks Signed-off-by: raver119 <raver119@gmail.com> * one more tweak Signed-off-by: raver119 <raver119@gmail.com> * - make float16 and bfloat16 to compile successfully on cuda Signed-off-by: Yurii <iuriish@yahoo.com> * - do not use resources of view-like arrays when move semantics is applied Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of pointers in signatures NDArray methods 1 Signed-off-by: Yurii <iuriish@yahoo.com> * - correction of signature of NDArray::dup method Signed-off-by: Yurii <iuriish@yahoo.com> * - correction of signature of NDArray::reduceAlongDimension method Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyIndexReduce and applyTrueBroadcast methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyReduce3 and varianceAlongDimension methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::tensorsAlongDimension and diagonal methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::allTensorsAlongDimension Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::reduceAlongDimension 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyTransform 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyPairwiseTransform 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyBroadcast 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyTrueBroadcast 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyScalar and applyScalarArr Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::lambda methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::reduce3 methods 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of following NDArray methods: add/sub/mul/div row/column and fillAsTriangular Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::tileToShape methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::isShapeSameStrict method Signed-off-by: Yurii <iuriish@yahoo.com> * minor corrections in tests Signed-off-by: Yurii <iuriish@yahoo.com> * - replace reduce op in batchnorm mkldnn Signed-off-by: Yurii <iuriish@yahoo.com> * - add explicit templates instantiations for operator+(NDArray&&. const scalar) Signed-off-by: Yurii <iuriish@yahoo.com> * - corrections of casts in float16/bfloat16 Signed-off-by: Yurii <iuriish@yahoo.com> * - provide move semantics in following NDArray methods: transform, applyTrueBroadcast, transpose, reshape, permute Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of input array A duplicate in svd cuda op Signed-off-by: Yurii <iuriish@yahoo.com> * - avoid available bug in svd cuda API Signed-off-by: Yurii <iuriish@yahoo.com> * - add temporary global memory buffer in svd cuda when calcUV = false and m != n Signed-off-by: Yurii <iuriish@yahoo.com> * - remove test with blfoat16 type for betainC Signed-off-by: Yurii <iuriish@yahoo.com> * - resolve conflicts after master has been merged in Signed-off-by: Yurii <iuriish@yahoo.com> * - changed type of affected input array in fused_batch_norm Signed-off-by: Yurii <iuriish@yahoo.com> * - add several explicit type castings Signed-off-by: Yurii <iuriish@yahoo.com> * - add ND4J_EXPORT to operators Signed-off-by: Yurii <iuriish@yahoo.com> * - add explicit template types in instantiations of template arithm operators of NDArray class Signed-off-by: Yurii <iuriish@yahoo.com> * - one more test fix Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com>
2019-12-20 20:35:39 +01:00
(dArrayBatch, reinterpret_cast<F *>(tempOutput.specialBuffer()), packX.specialOffsets(), batchSize);
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
status = cusolverDnSetStream(handle, *stream);
if (CUSOLVER_STATUS_SUCCESS != status) {
throw cuda_exception::build("helpers::cholesky_: Cannot set stream to solver handle", status);
}
const cublasFillMode_t uplo = CUBLAS_FILL_MODE_UPPER;
if (input->dataType() == DataType::DOUBLE)
status = cusolverDnDpotrfBatched(
handle,
uplo,
n,
(double **) dArrayBatch,
n,
dInfoArray,
batchSize);
else
status = cusolverDnSpotrfBatched(
handle,
uplo,
n,
(float **) dArrayBatch,
n,
dInfoArray,
batchSize);
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
if (CUSOLVER_STATUS_SUCCESS != status) {
throw cuda_exception::build("helpers::cholesky_: Cholesky factorization failed for batch", status);
}
adjustResultsKernel<F> << < batchSize, n2, 128, *stream >> >
Shyrma temp (#131) * - specifying template instantiation for certain types in float16 and bloat16 Signed-off-by: Yurii <iuriish@yahoo.com> * - polishing bfloat16 and float16 member functions template specialization Signed-off-by: Yurii <iuriish@yahoo.com> * - rewrite and overload array +-*/ scalar and scalar +-*/ arr in NDAray class Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections which have to do with and rvalue lvalue conversions Signed-off-by: Yurii <iuriish@yahoo.com> * - provide move semantic in NDArray operators array +-/* array Signed-off-by: Yurii <iuriish@yahoo.com> * float16/bfloat16 tweaks Signed-off-by: raver119 <raver119@gmail.com> * one more tweak Signed-off-by: raver119 <raver119@gmail.com> * - make float16 and bfloat16 to compile successfully on cuda Signed-off-by: Yurii <iuriish@yahoo.com> * - do not use resources of view-like arrays when move semantics is applied Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of pointers in signatures NDArray methods 1 Signed-off-by: Yurii <iuriish@yahoo.com> * - correction of signature of NDArray::dup method Signed-off-by: Yurii <iuriish@yahoo.com> * - correction of signature of NDArray::reduceAlongDimension method Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyIndexReduce and applyTrueBroadcast methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyReduce3 and varianceAlongDimension methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::tensorsAlongDimension and diagonal methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::allTensorsAlongDimension Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::reduceAlongDimension 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyTransform 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyPairwiseTransform 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyBroadcast 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyTrueBroadcast 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyScalar and applyScalarArr Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::lambda methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::reduce3 methods 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of following NDArray methods: add/sub/mul/div row/column and fillAsTriangular Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::tileToShape methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::isShapeSameStrict method Signed-off-by: Yurii <iuriish@yahoo.com> * minor corrections in tests Signed-off-by: Yurii <iuriish@yahoo.com> * - replace reduce op in batchnorm mkldnn Signed-off-by: Yurii <iuriish@yahoo.com> * - add explicit templates instantiations for operator+(NDArray&&. const scalar) Signed-off-by: Yurii <iuriish@yahoo.com> * - corrections of casts in float16/bfloat16 Signed-off-by: Yurii <iuriish@yahoo.com> * - provide move semantics in following NDArray methods: transform, applyTrueBroadcast, transpose, reshape, permute Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of input array A duplicate in svd cuda op Signed-off-by: Yurii <iuriish@yahoo.com> * - avoid available bug in svd cuda API Signed-off-by: Yurii <iuriish@yahoo.com> * - add temporary global memory buffer in svd cuda when calcUV = false and m != n Signed-off-by: Yurii <iuriish@yahoo.com> * - remove test with blfoat16 type for betainC Signed-off-by: Yurii <iuriish@yahoo.com> * - resolve conflicts after master has been merged in Signed-off-by: Yurii <iuriish@yahoo.com> * - changed type of affected input array in fused_batch_norm Signed-off-by: Yurii <iuriish@yahoo.com> * - add several explicit type castings Signed-off-by: Yurii <iuriish@yahoo.com> * - add ND4J_EXPORT to operators Signed-off-by: Yurii <iuriish@yahoo.com> * - add explicit template types in instantiations of template arithm operators of NDArray class Signed-off-by: Yurii <iuriish@yahoo.com> * - one more test fix Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com>
2019-12-20 20:35:39 +01:00
(reinterpret_cast<F *>(tempOutput.specialBuffer()), packX.specialShapeInfo(), packX.specialOffsets(), batchSize, n);
err = cudaFree(dArrayBatch);
if (err) {
throw cuda_exception::build("helpers::cholesky_: Cannot deallocate memory for solver batch data buffer",
err);
}
err = cudaFree(dInfoArray);
if (err) {
throw cuda_exception::build("helpers::cholesky_: Cannot allocate memory for solver errors buffer", err);
}
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
if (!inplace)
Shyrma temp (#131) * - specifying template instantiation for certain types in float16 and bloat16 Signed-off-by: Yurii <iuriish@yahoo.com> * - polishing bfloat16 and float16 member functions template specialization Signed-off-by: Yurii <iuriish@yahoo.com> * - rewrite and overload array +-*/ scalar and scalar +-*/ arr in NDAray class Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections which have to do with and rvalue lvalue conversions Signed-off-by: Yurii <iuriish@yahoo.com> * - provide move semantic in NDArray operators array +-/* array Signed-off-by: Yurii <iuriish@yahoo.com> * float16/bfloat16 tweaks Signed-off-by: raver119 <raver119@gmail.com> * one more tweak Signed-off-by: raver119 <raver119@gmail.com> * - make float16 and bfloat16 to compile successfully on cuda Signed-off-by: Yurii <iuriish@yahoo.com> * - do not use resources of view-like arrays when move semantics is applied Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of pointers in signatures NDArray methods 1 Signed-off-by: Yurii <iuriish@yahoo.com> * - correction of signature of NDArray::dup method Signed-off-by: Yurii <iuriish@yahoo.com> * - correction of signature of NDArray::reduceAlongDimension method Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyIndexReduce and applyTrueBroadcast methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyReduce3 and varianceAlongDimension methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::tensorsAlongDimension and diagonal methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::allTensorsAlongDimension Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::reduceAlongDimension 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyTransform 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyPairwiseTransform 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyBroadcast 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyTrueBroadcast 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyScalar and applyScalarArr Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::lambda methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::reduce3 methods 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of following NDArray methods: add/sub/mul/div row/column and fillAsTriangular Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::tileToShape methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::isShapeSameStrict method Signed-off-by: Yurii <iuriish@yahoo.com> * minor corrections in tests Signed-off-by: Yurii <iuriish@yahoo.com> * - replace reduce op in batchnorm mkldnn Signed-off-by: Yurii <iuriish@yahoo.com> * - add explicit templates instantiations for operator+(NDArray&&. const scalar) Signed-off-by: Yurii <iuriish@yahoo.com> * - corrections of casts in float16/bfloat16 Signed-off-by: Yurii <iuriish@yahoo.com> * - provide move semantics in following NDArray methods: transform, applyTrueBroadcast, transpose, reshape, permute Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of input array A duplicate in svd cuda op Signed-off-by: Yurii <iuriish@yahoo.com> * - avoid available bug in svd cuda API Signed-off-by: Yurii <iuriish@yahoo.com> * - add temporary global memory buffer in svd cuda when calcUV = false and m != n Signed-off-by: Yurii <iuriish@yahoo.com> * - remove test with blfoat16 type for betainC Signed-off-by: Yurii <iuriish@yahoo.com> * - resolve conflicts after master has been merged in Signed-off-by: Yurii <iuriish@yahoo.com> * - changed type of affected input array in fused_batch_norm Signed-off-by: Yurii <iuriish@yahoo.com> * - add several explicit type castings Signed-off-by: Yurii <iuriish@yahoo.com> * - add ND4J_EXPORT to operators Signed-off-by: Yurii <iuriish@yahoo.com> * - add explicit template types in instantiations of template arithm operators of NDArray class Signed-off-by: Yurii <iuriish@yahoo.com> * - one more test fix Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com>
2019-12-20 20:35:39 +01:00
output->assign(tempOutput);
else
Shyrma temp (#131) * - specifying template instantiation for certain types in float16 and bloat16 Signed-off-by: Yurii <iuriish@yahoo.com> * - polishing bfloat16 and float16 member functions template specialization Signed-off-by: Yurii <iuriish@yahoo.com> * - rewrite and overload array +-*/ scalar and scalar +-*/ arr in NDAray class Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections which have to do with and rvalue lvalue conversions Signed-off-by: Yurii <iuriish@yahoo.com> * - provide move semantic in NDArray operators array +-/* array Signed-off-by: Yurii <iuriish@yahoo.com> * float16/bfloat16 tweaks Signed-off-by: raver119 <raver119@gmail.com> * one more tweak Signed-off-by: raver119 <raver119@gmail.com> * - make float16 and bfloat16 to compile successfully on cuda Signed-off-by: Yurii <iuriish@yahoo.com> * - do not use resources of view-like arrays when move semantics is applied Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of pointers in signatures NDArray methods 1 Signed-off-by: Yurii <iuriish@yahoo.com> * - correction of signature of NDArray::dup method Signed-off-by: Yurii <iuriish@yahoo.com> * - correction of signature of NDArray::reduceAlongDimension method Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyIndexReduce and applyTrueBroadcast methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyReduce3 and varianceAlongDimension methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::tensorsAlongDimension and diagonal methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::allTensorsAlongDimension Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::reduceAlongDimension 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyTransform 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyPairwiseTransform 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyBroadcast 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyTrueBroadcast 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyScalar and applyScalarArr Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::lambda methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::reduce3 methods 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of following NDArray methods: add/sub/mul/div row/column and fillAsTriangular Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::tileToShape methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::isShapeSameStrict method Signed-off-by: Yurii <iuriish@yahoo.com> * minor corrections in tests Signed-off-by: Yurii <iuriish@yahoo.com> * - replace reduce op in batchnorm mkldnn Signed-off-by: Yurii <iuriish@yahoo.com> * - add explicit templates instantiations for operator+(NDArray&&. const scalar) Signed-off-by: Yurii <iuriish@yahoo.com> * - corrections of casts in float16/bfloat16 Signed-off-by: Yurii <iuriish@yahoo.com> * - provide move semantics in following NDArray methods: transform, applyTrueBroadcast, transpose, reshape, permute Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of input array A duplicate in svd cuda op Signed-off-by: Yurii <iuriish@yahoo.com> * - avoid available bug in svd cuda API Signed-off-by: Yurii <iuriish@yahoo.com> * - add temporary global memory buffer in svd cuda when calcUV = false and m != n Signed-off-by: Yurii <iuriish@yahoo.com> * - remove test with blfoat16 type for betainC Signed-off-by: Yurii <iuriish@yahoo.com> * - resolve conflicts after master has been merged in Signed-off-by: Yurii <iuriish@yahoo.com> * - changed type of affected input array in fused_batch_norm Signed-off-by: Yurii <iuriish@yahoo.com> * - add several explicit type castings Signed-off-by: Yurii <iuriish@yahoo.com> * - add ND4J_EXPORT to operators Signed-off-by: Yurii <iuriish@yahoo.com> * - add explicit template types in instantiations of template arithm operators of NDArray class Signed-off-by: Yurii <iuriish@yahoo.com> * - one more test fix Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com>
2019-12-20 20:35:39 +01:00
input->assign(tempOutput);
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
NDArray::registerSpecialUse({output}, {input});
return Status::OK();
}
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
// template <typename T>
int cholesky_(LaunchContext *context, NDArray *input, NDArray *output, bool inplace) {
NDArray::prepareSpecialUse({output}, {input});
if (input->dataType() == DataType::DOUBLE)
cholesky__<double>(context, input, output, inplace);
else if (input->dataType() == DataType::FLOAT32)
cholesky__<float>(context, input, output, inplace);
else {
std::unique_ptr<NDArray> tempOutput(
NDArrayFactory::create_('c', input->getShapeAsVector(), DataType::FLOAT32, context));
tempOutput->assign(input);
cholesky__<float>(context, tempOutput.get(), tempOutput.get(), true);
output->assign(tempOutput.get());
}
NDArray::registerSpecialUse({output}, {input});
return Status::OK();
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
2019-06-06 14:21:15 +02:00
int cholesky(sd::LaunchContext *context, NDArray *input, NDArray *output, bool inplace) {
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
// BUILD_SINGLE_SELECTOR(input->dataType(), return cholesky_, (context, input, output, inplace), FLOAT_TYPES);
return cholesky_(context, input, output, inplace);
}
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
// BUILD_SINGLE_TEMPLATE(template int cholesky_, (LaunchContext* context, NDArray* input, NDArray* output, bool inplace), FLOAT_TYPES);
BUILD_SINGLE_TEMPLATE(template int inverse_, (sd::LaunchContext * context, NDArray * input, NDArray * output),
FLOAT_NATIVE);
2019-06-06 14:21:15 +02:00
template<typename T>
__global__ void
logDetKernel(T *inputBuf, Nd4jLong *inputShape, Nd4jLong batchNum, Nd4jLong *tadShape, Nd4jLong *tadOffsets,
T *outputBuf, Nd4jLong *outputShape) {
[WIP] More of CUDA (#95) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Implementation of hashcode cuda helper. Working edition. * Fixed parallel test input arangements. * Fixed tests for hashcode op. * Fixed shape calculation for image:crop_and_resize op and test. * NativeOps tests. Initial test suite. * Added tests for indexReduce methods. * Added test on execBroadcast with NDArray as dimensions. * Added test on execBroadcastBool with NDArray as dimensions. * Added tests on execPairwiseTransform and execPairwiseTransofrmBool. * Added tests for execReduce with scalar results. * Added reduce tests for non-empty dims array. * Added tests for reduce3. * Added tests for execScalar. * Added tests for execSummaryStats. * - provide cpu/cuda code for batch_to_space - testing it Signed-off-by: Yurii <yurii@skymind.io> * - remove old test for batch_to_space (had wrong format and numbers were not checked) Signed-off-by: Yurii <yurii@skymind.io> * Fixed complilation errors with test. * Added test for execTransformFloat. * Added test for execTransformSame. * Added test for execTransformBool. * Added test for execTransformStrict. * Added tests for execScalar/execScalarBool with TADs. * Added test for flatten. * - provide cpu/cuda code for space_to_Batch operaion Signed-off-by: Yurii <yurii@skymind.io> * Added test for concat. * comment unnecessary stuff in s_t_b Signed-off-by: Yurii <yurii@skymind.io> * Added test for specialConcat. * Added tests for memcpy/set routines. * Fixed pullRow cuda test. * Added pullRow test. * Added average test. * - correct typo in NDArray::applyPairwiseTransform(nd4j::pairwise::BoolOps op...) Signed-off-by: Yurii <yurii@skymind.io> * - debugging and fixing cuda tests in JavaInteropTests file Signed-off-by: Yurii <yurii@skymind.io> * - correct some tests Signed-off-by: Yurii <yurii@skymind.io> * Added test for shuffle. * Fixed ops declarations. * Restored omp and added shuffle test. * Added convertTypes test. * Added tests for execRandom. Eliminated usage of RandomBuffer with NativeOps. * Added sort tests. * Added tests for execCustomOp. * - further debuging and fixing tests terminated with crash Signed-off-by: Yurii <yurii@skymind.io> * Added tests for calculateOutputShapes. * Addded Benchmarks test. * Commented benchmark tests. * change assertion Signed-off-by: raver119 <raver119@gmail.com> * Added tests for apply_sgd op. Added cpu helper for that op. * Implement cuda helper for aplly_sgd op. Fixed tests for NativeOps. * Added test for assign broadcastable. * Added tests for assign_bp op. * Added tests for axpy op. * - assign/execScalar/execTransformAny signature change - minor test fix Signed-off-by: raver119 <raver119@gmail.com> * Fixed axpy op. * meh Signed-off-by: raver119 <raver119@gmail.com> * - fix tests for nativeOps::concat Signed-off-by: Yurii <yurii@skymind.io> * sequential transform/scalar Signed-off-by: raver119 <raver119@gmail.com> * allow nested parallelism Signed-off-by: raver119 <raver119@gmail.com> * assign_bp leak fix Signed-off-by: raver119 <raver119@gmail.com> * block setRNG fix Signed-off-by: raver119 <raver119@gmail.com> * enable parallelism by default Signed-off-by: raver119 <raver119@gmail.com> * enable nested parallelism by default Signed-off-by: raver119 <raver119@gmail.com> * Added cuda implementation for row_count helper. * Added implementation for tnse gains op helper. * - take into account possible situations when input arrays are empty in reduce_ cuda stuff Signed-off-by: Yurii <yurii@skymind.io> * Implemented tsne/edge_forces op cuda-based helper. Parallelized cpu-based helper for edge_forces. * Added kernel for tsne/symmetrized op heleper. * Implementation of tsne/symmetrized op cuda helper. Working edition. * Eliminated waste printfs. * Added test for broadcastgradientargs op. * host-only fallback for empty reduce float Signed-off-by: raver119 <raver119@gmail.com> * - some tests fixes Signed-off-by: Yurii <yurii@skymind.io> * - correct the rest of reduce_ stuff Signed-off-by: Yurii <yurii@skymind.io> * - further correction of reduce_ stuff Signed-off-by: Yurii <yurii@skymind.io> * Added test for Cbow op. Also added cuda implementation for cbow helpers. * - improve code of stack operation for scalar case Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda kernel for gatherND operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cbow helpers with cuda kernels. * minor tests tweaks Signed-off-by: raver119 <raver119@gmail.com> * minor tests tweaks Signed-off-by: raver119 <raver119@gmail.com> * - further correction of cuda stuff Signed-off-by: Yurii <yurii@skymind.io> * Implementatation of cbow op helper with cuda kernels. Working edition. * Skip random testing for cudablas case. * lstmBlockCell context fix Signed-off-by: raver119 <raver119@gmail.com> * Added tests for ELU and ELU_BP ops. * Added tests for eq_scalar, gt_scalar, gte_scalar and lte_scalar ops. * Added tests for neq_scalar. * Added test for noop. * - further work on clipbynorm_bp Signed-off-by: Yurii <yurii@skymind.io> * - get rid of concat op call, use instead direct concat helper call Signed-off-by: Yurii <yurii@skymind.io> * lstmBlockCell context fix Signed-off-by: raver119 <raver119@gmail.com> * Added tests for lrelu and lrelu_bp. * Added tests for selu and selu_bp. * Fixed lrelu derivative helpers. * - some corrections in lstm Signed-off-by: Yurii <yurii@skymind.io> * operator * result shape fix Signed-off-by: raver119 <raver119@gmail.com> * - correct typo in lstmCell Signed-off-by: Yurii <yurii@skymind.io> * few tests fixed Signed-off-by: raver119 <raver119@gmail.com> * CUDA inverse broadcast bool fix Signed-off-by: raver119 <raver119@gmail.com> * disable MMAP test for CUDA Signed-off-by: raver119 <raver119@gmail.com> * BooleanOp syncToDevice Signed-off-by: raver119 <raver119@gmail.com> * meh Signed-off-by: raver119 <raver119@gmail.com> * additional data types for im2col/col2im Signed-off-by: raver119 <raver119@gmail.com> * Added test for firas_sparse op. * one more RandomBuffer test excluded Signed-off-by: raver119 <raver119@gmail.com> * Added tests for flatten op. * Added test for Floor op. * bunch of tests fixed Signed-off-by: raver119 <raver119@gmail.com> * mmulDot tests fixed Signed-off-by: raver119 <raver119@gmail.com> * more tests fixed Signed-off-by: raver119 <raver119@gmail.com> * Implemented floordiv_bp op and tests. * Fixed scalar case with cuda implementation for bds. * - work on cuda kernel for clip_by_norm backprop op is completed Signed-off-by: Yurii <yurii@skymind.io> * Eliminate cbow crach. * more tests fixed Signed-off-by: raver119 <raver119@gmail.com> * more tests fixed Signed-off-by: raver119 <raver119@gmail.com> * Eliminated abortion with batched nlp test. * more tests fixed Signed-off-by: raver119 <raver119@gmail.com> * Fixed shared flag initializing. * disabled bunch of cpu workspaces tests Signed-off-by: raver119 <raver119@gmail.com> * scalar operators fix: missing registerSpecialUse call Signed-off-by: raver119 <raver119@gmail.com> * Fixed logdet for cuda and tests. * - correct clipBynorm_bp Signed-off-by: Yurii <yurii@skymind.io> * Fixed crop_and_resize shape datatype. * - correct some mmul tests Signed-off-by: Yurii <yurii@skymind.io>
2019-08-02 19:01:03 +02:00
__shared__ int n;
if (threadIdx.x == 0) {
n = shape::sizeAt(inputShape, -1); // * shape::sizeAt(inputShape, -1);
}
__syncthreads();
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
T *output = outputBuf;
T *input = inputBuf;
[WIP] More of CUDA (#95) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Implementation of hashcode cuda helper. Working edition. * Fixed parallel test input arangements. * Fixed tests for hashcode op. * Fixed shape calculation for image:crop_and_resize op and test. * NativeOps tests. Initial test suite. * Added tests for indexReduce methods. * Added test on execBroadcast with NDArray as dimensions. * Added test on execBroadcastBool with NDArray as dimensions. * Added tests on execPairwiseTransform and execPairwiseTransofrmBool. * Added tests for execReduce with scalar results. * Added reduce tests for non-empty dims array. * Added tests for reduce3. * Added tests for execScalar. * Added tests for execSummaryStats. * - provide cpu/cuda code for batch_to_space - testing it Signed-off-by: Yurii <yurii@skymind.io> * - remove old test for batch_to_space (had wrong format and numbers were not checked) Signed-off-by: Yurii <yurii@skymind.io> * Fixed complilation errors with test. * Added test for execTransformFloat. * Added test for execTransformSame. * Added test for execTransformBool. * Added test for execTransformStrict. * Added tests for execScalar/execScalarBool with TADs. * Added test for flatten. * - provide cpu/cuda code for space_to_Batch operaion Signed-off-by: Yurii <yurii@skymind.io> * Added test for concat. * comment unnecessary stuff in s_t_b Signed-off-by: Yurii <yurii@skymind.io> * Added test for specialConcat. * Added tests for memcpy/set routines. * Fixed pullRow cuda test. * Added pullRow test. * Added average test. * - correct typo in NDArray::applyPairwiseTransform(nd4j::pairwise::BoolOps op...) Signed-off-by: Yurii <yurii@skymind.io> * - debugging and fixing cuda tests in JavaInteropTests file Signed-off-by: Yurii <yurii@skymind.io> * - correct some tests Signed-off-by: Yurii <yurii@skymind.io> * Added test for shuffle. * Fixed ops declarations. * Restored omp and added shuffle test. * Added convertTypes test. * Added tests for execRandom. Eliminated usage of RandomBuffer with NativeOps. * Added sort tests. * Added tests for execCustomOp. * - further debuging and fixing tests terminated with crash Signed-off-by: Yurii <yurii@skymind.io> * Added tests for calculateOutputShapes. * Addded Benchmarks test. * Commented benchmark tests. * change assertion Signed-off-by: raver119 <raver119@gmail.com> * Added tests for apply_sgd op. Added cpu helper for that op. * Implement cuda helper for aplly_sgd op. Fixed tests for NativeOps. * Added test for assign broadcastable. * Added tests for assign_bp op. * Added tests for axpy op. * - assign/execScalar/execTransformAny signature change - minor test fix Signed-off-by: raver119 <raver119@gmail.com> * Fixed axpy op. * meh Signed-off-by: raver119 <raver119@gmail.com> * - fix tests for nativeOps::concat Signed-off-by: Yurii <yurii@skymind.io> * sequential transform/scalar Signed-off-by: raver119 <raver119@gmail.com> * allow nested parallelism Signed-off-by: raver119 <raver119@gmail.com> * assign_bp leak fix Signed-off-by: raver119 <raver119@gmail.com> * block setRNG fix Signed-off-by: raver119 <raver119@gmail.com> * enable parallelism by default Signed-off-by: raver119 <raver119@gmail.com> * enable nested parallelism by default Signed-off-by: raver119 <raver119@gmail.com> * Added cuda implementation for row_count helper. * Added implementation for tnse gains op helper. * - take into account possible situations when input arrays are empty in reduce_ cuda stuff Signed-off-by: Yurii <yurii@skymind.io> * Implemented tsne/edge_forces op cuda-based helper. Parallelized cpu-based helper for edge_forces. * Added kernel for tsne/symmetrized op heleper. * Implementation of tsne/symmetrized op cuda helper. Working edition. * Eliminated waste printfs. * Added test for broadcastgradientargs op. * host-only fallback for empty reduce float Signed-off-by: raver119 <raver119@gmail.com> * - some tests fixes Signed-off-by: Yurii <yurii@skymind.io> * - correct the rest of reduce_ stuff Signed-off-by: Yurii <yurii@skymind.io> * - further correction of reduce_ stuff Signed-off-by: Yurii <yurii@skymind.io> * Added test for Cbow op. Also added cuda implementation for cbow helpers. * - improve code of stack operation for scalar case Signed-off-by: Yurii <yurii@skymind.io> * - provide cuda kernel for gatherND operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cbow helpers with cuda kernels. * minor tests tweaks Signed-off-by: raver119 <raver119@gmail.com> * minor tests tweaks Signed-off-by: raver119 <raver119@gmail.com> * - further correction of cuda stuff Signed-off-by: Yurii <yurii@skymind.io> * Implementatation of cbow op helper with cuda kernels. Working edition. * Skip random testing for cudablas case. * lstmBlockCell context fix Signed-off-by: raver119 <raver119@gmail.com> * Added tests for ELU and ELU_BP ops. * Added tests for eq_scalar, gt_scalar, gte_scalar and lte_scalar ops. * Added tests for neq_scalar. * Added test for noop. * - further work on clipbynorm_bp Signed-off-by: Yurii <yurii@skymind.io> * - get rid of concat op call, use instead direct concat helper call Signed-off-by: Yurii <yurii@skymind.io> * lstmBlockCell context fix Signed-off-by: raver119 <raver119@gmail.com> * Added tests for lrelu and lrelu_bp. * Added tests for selu and selu_bp. * Fixed lrelu derivative helpers. * - some corrections in lstm Signed-off-by: Yurii <yurii@skymind.io> * operator * result shape fix Signed-off-by: raver119 <raver119@gmail.com> * - correct typo in lstmCell Signed-off-by: Yurii <yurii@skymind.io> * few tests fixed Signed-off-by: raver119 <raver119@gmail.com> * CUDA inverse broadcast bool fix Signed-off-by: raver119 <raver119@gmail.com> * disable MMAP test for CUDA Signed-off-by: raver119 <raver119@gmail.com> * BooleanOp syncToDevice Signed-off-by: raver119 <raver119@gmail.com> * meh Signed-off-by: raver119 <raver119@gmail.com> * additional data types for im2col/col2im Signed-off-by: raver119 <raver119@gmail.com> * Added test for firas_sparse op. * one more RandomBuffer test excluded Signed-off-by: raver119 <raver119@gmail.com> * Added tests for flatten op. * Added test for Floor op. * bunch of tests fixed Signed-off-by: raver119 <raver119@gmail.com> * mmulDot tests fixed Signed-off-by: raver119 <raver119@gmail.com> * more tests fixed Signed-off-by: raver119 <raver119@gmail.com> * Implemented floordiv_bp op and tests. * Fixed scalar case with cuda implementation for bds. * - work on cuda kernel for clip_by_norm backprop op is completed Signed-off-by: Yurii <yurii@skymind.io> * Eliminate cbow crach. * more tests fixed Signed-off-by: raver119 <raver119@gmail.com> * more tests fixed Signed-off-by: raver119 <raver119@gmail.com> * Eliminated abortion with batched nlp test. * more tests fixed Signed-off-by: raver119 <raver119@gmail.com> * Fixed shared flag initializing. * disabled bunch of cpu workspaces tests Signed-off-by: raver119 <raver119@gmail.com> * scalar operators fix: missing registerSpecialUse call Signed-off-by: raver119 <raver119@gmail.com> * Fixed logdet for cuda and tests. * - correct clipBynorm_bp Signed-off-by: Yurii <yurii@skymind.io> * Fixed crop_and_resize shape datatype. * - correct some mmul tests Signed-off-by: Yurii <yurii@skymind.io>
2019-08-02 19:01:03 +02:00
for (auto i = blockIdx.x; i < batchNum; i += gridDim.x) {
T *current = input + tadOffsets[i];
auto zIndex = shape::getIndexOffset(i, outputShape);
for (auto e = threadIdx.x; e < n; e += blockDim.x) {
Nd4jLong diag[] = {e, e};
auto xIndex = shape::getOffset(tadShape, diag);
math::atomics::nd4j_atomicAdd(&output[zIndex],math::nd4j_log<T, T>(current[xIndex] * current[xIndex]));
}
[WIP] more CUDA stuff (#57) * initial commit Signed-off-by: raver119 <raver119@gmail.com> * Added gradcheck test for dynamic_partition_bp op. * - implementation of dilation op (cpu and cuda) Signed-off-by: Yurii <yurii@skymind.io> * Fixed broadcast_dynamic_shape 1D case and tests. * Fixed usage of default integer arguments. * Fixed dynamic_partition_bp op and tests. * Eliminated test with grad check for dynamic_partition_bp op. * start working on cuda svd - porting available corresponding api from cuSOLVER library Signed-off-by: Yurii <yurii@skymind.io> * provide prelu_bp Signed-off-by: Yurii <yurii@skymind.io> * - provide gruCell_bp (old version ??) Signed-off-by: Yurii <yurii@skymind.io> * - polishing cumsum_bp and cumprod_bp tests Signed-off-by: Yurii <yurii@skymind.io> * provide sparseSoftmaxCrossEntropyWithLogits and sparseSoftmaxCrossEntropyWithLogits_grad Signed-off-by: Yurii <yurii@skymind.io> * Fixed atomicMul with float input/output * implementation of cuda kernel for triu_bp operation Signed-off-by: Yurii <yurii@skymind.io> * Refactored lup helper to add parrallel computing. * cusolver libraries Signed-off-by: raver119 <raver119@gmail.com> * uncomment cuSolver APIs in svd.cu Signed-off-by: Yurii <yurii@skymind.io> * cusolver var Signed-off-by: raver119 <raver119@gmail.com> * - further work on cuSolver svd Signed-off-by: Yurii <yurii@skymind.io> * Implement usage of cuda solver to LUP decomposition. * - correct naames in lup functions Signed-off-by: Yurii <yurii@skymind.io> * correct svdQR cuda Signed-off-by: Yurii <yurii@skymind.io> * - provide transpositions of input matrices in case of c order in svdCudaQR Signed-off-by: Yurii <yurii@skymind.io> * Fixed implementation issues with LUP usign cuda solver. * Implementation of matrix_determinant helper with cuda kernels. Working revision. * Implemented log_matrix_determinant helper with cuda kernels. * - implementation of batched cuda svd Signed-off-by: Yurii <yurii@skymind.io> * Refactored cholesky helper and implementation of cuda solver cholesky batch. * - implementation of cuda kernel for tile bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of cholesky and logdet with cuda kernels. * - implementation of cuda kernel for sru_bidirectional Signed-off-by: Yurii <yurii@skymind.io> * Fixed cholesky helper. * Cholesky op helper implementation. Working double-based cublas implementation. * bad import excluded Signed-off-by: raver119 <raver119@gmail.com> * Finished with cuda implementation of cholesky helper and tests. * - implementation of cuda kernel for sru_bidirectional_backprop operation Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse op helper with cuda kernels. The first revision. * - start working on gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * Implementation of matrix_inverse helper. * - further work on new gruCell_bp Signed-off-by: Yurii <yurii@skymind.io> * cuBLAS related fixes Signed-off-by: raver119 <raver119@gmail.com> * calculateOutputShapes() now passes device buffers as well Signed-off-by: raver119 <raver119@gmail.com> * special concat/average/accumulate init host pointers now Signed-off-by: raver119 <raver119@gmail.com> * few more tweaks Signed-off-by: raver119 <raver119@gmail.com> * additional CudaDataBufferFactory signatures certain for data types Signed-off-by: raver119 <raver119@gmail.com> * cuSolver host buffer Signed-off-by: raver119 <raver119@gmail.com> * buffer to buffer memcpy host ptr allocation Signed-off-by: raver119 <raver119@gmail.com>
2019-07-12 10:51:51 +02:00
}
}
2019-06-06 14:21:15 +02:00
template<typename T>
int logdetFunctor_(sd::LaunchContext *context, NDArray *input, NDArray *output) {
NDArray::prepareSpecialUse({output}, {input});
auto n2 = input->sizeAt(-1) * input->sizeAt(-2);
auto stream = context->getCudaStream();
NDArray tempOutput(*input);
cholesky(context, input, &tempOutput, false);
auto outputBuf = output->dataBuffer()->specialAsT<T>(); //reinterpret_cast<T*>(output->specialBuffer()); // + e * n2; // + e * n2;
Shyrma temp (#131) * - specifying template instantiation for certain types in float16 and bloat16 Signed-off-by: Yurii <iuriish@yahoo.com> * - polishing bfloat16 and float16 member functions template specialization Signed-off-by: Yurii <iuriish@yahoo.com> * - rewrite and overload array +-*/ scalar and scalar +-*/ arr in NDAray class Signed-off-by: Yurii <iuriish@yahoo.com> * - make corrections which have to do with and rvalue lvalue conversions Signed-off-by: Yurii <iuriish@yahoo.com> * - provide move semantic in NDArray operators array +-/* array Signed-off-by: Yurii <iuriish@yahoo.com> * float16/bfloat16 tweaks Signed-off-by: raver119 <raver119@gmail.com> * one more tweak Signed-off-by: raver119 <raver119@gmail.com> * - make float16 and bfloat16 to compile successfully on cuda Signed-off-by: Yurii <iuriish@yahoo.com> * - do not use resources of view-like arrays when move semantics is applied Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of pointers in signatures NDArray methods 1 Signed-off-by: Yurii <iuriish@yahoo.com> * - correction of signature of NDArray::dup method Signed-off-by: Yurii <iuriish@yahoo.com> * - correction of signature of NDArray::reduceAlongDimension method Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyIndexReduce and applyTrueBroadcast methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyReduce3 and varianceAlongDimension methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::tensorsAlongDimension and diagonal methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::allTensorsAlongDimension Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::reduceAlongDimension 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyTransform 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyPairwiseTransform 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyBroadcast 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyTrueBroadcast 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::applyScalar and applyScalarArr Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::lambda methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::reduce3 methods 2 Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of following NDArray methods: add/sub/mul/div row/column and fillAsTriangular Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::tileToShape methods Signed-off-by: Yurii <iuriish@yahoo.com> * - signature correction of NDArray::isShapeSameStrict method Signed-off-by: Yurii <iuriish@yahoo.com> * minor corrections in tests Signed-off-by: Yurii <iuriish@yahoo.com> * - replace reduce op in batchnorm mkldnn Signed-off-by: Yurii <iuriish@yahoo.com> * - add explicit templates instantiations for operator+(NDArray&&. const scalar) Signed-off-by: Yurii <iuriish@yahoo.com> * - corrections of casts in float16/bfloat16 Signed-off-by: Yurii <iuriish@yahoo.com> * - provide move semantics in following NDArray methods: transform, applyTrueBroadcast, transpose, reshape, permute Signed-off-by: Yurii <iuriish@yahoo.com> * - get rid of input array A duplicate in svd cuda op Signed-off-by: Yurii <iuriish@yahoo.com> * - avoid available bug in svd cuda API Signed-off-by: Yurii <iuriish@yahoo.com> * - add temporary global memory buffer in svd cuda when calcUV = false and m != n Signed-off-by: Yurii <iuriish@yahoo.com> * - remove test with blfoat16 type for betainC Signed-off-by: Yurii <iuriish@yahoo.com> * - resolve conflicts after master has been merged in Signed-off-by: Yurii <iuriish@yahoo.com> * - changed type of affected input array in fused_batch_norm Signed-off-by: Yurii <iuriish@yahoo.com> * - add several explicit type castings Signed-off-by: Yurii <iuriish@yahoo.com> * - add ND4J_EXPORT to operators Signed-off-by: Yurii <iuriish@yahoo.com> * - add explicit template types in instantiations of template arithm operators of NDArray class Signed-off-by: Yurii <iuriish@yahoo.com> * - one more test fix Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: raver119 <raver119@gmail.com>
2019-12-20 20:35:39 +01:00
auto inputBuf = tempOutput.dataBuffer()->specialAsT<T>(); //reinterpret_cast<T*>(tempOutput.specialBuffer());
output->nullify();
auto packX = sd::ConstantTadHelper::getInstance()->tadForDimensions(tempOutput.getShapeInfo(),
{tempOutput.rankOf() - 2,
tempOutput.rankOf() - 1});
logDetKernel<T> <<<128, 512, 256, *stream>>>(inputBuf, tempOutput.specialShapeInfo(),
packX.numberOfTads(), packX.specialShapeInfo(),
packX.specialOffsets(), outputBuf, output->specialShapeInfo());
output->tickWriteDevice();
NDArray::registerSpecialUse({output}, {input});
return Status::OK();
}
int logdetFunctor(sd::LaunchContext *context, NDArray *input, NDArray *output) {
BUILD_SINGLE_SELECTOR(output->dataType(), return logdetFunctor_, (context, input, output), FLOAT_NATIVE);
}
/*
* lup - batched input, batched outputs
* */
int lup(LaunchContext *context, NDArray *input, NDArray *compound, NDArray *permutation) {
BUILD_DOUBLE_SELECTOR(input->dataType(), permutation->dataType(), lup_,(context, input, compound, permutation), FLOAT_NATIVE, INDEXING_TYPES);
return Status::OK();
}
// BUILD_SINGLE_TEMPLATE(template int logdetFunctor_,
// (sd::LaunchContext * context, NDArray * input, NDArray * output), FLOAT_NATIVE);
2019-06-06 14:21:15 +02:00
}
}
}