cavis/libnd4j/include/loops/cuda/transform/transform_same.cu

124 lines
4.9 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 <system/Environment.h>
2019-06-06 14:21:15 +02:00
#include <loops/transform_same.h>
#include <types/types.h>
#include <system/op_boilerplate.h>
2019-06-06 14:21:15 +02:00
#include <loops/legacy_ops.h>
#include <helpers/DebugHelper.h>
using namespace simdOps;
template <typename X, typename OpType>
__global__ void transformSameSimple(void *x, Nd4jLong *xShapeInfo, int xRank,
void *params,
void *z, Nd4jLong *zShapeInfo, int zRank,
int *allocationPointer,
void *reductionPointer,
Nd4jLong *tadShapeInfo, Nd4jLong *tadOffsets) {
functions::transform::TransformSame<X>::template transformCuda<OpType>(x,xShapeInfo,params,z,zShapeInfo,allocationPointer,reductionPointer, tadShapeInfo, tadOffsets);
}
namespace functions {
namespace transform {
template<typename X>
_CUDA_H void TransformSame<X>::executeTransformShaped(dim3 launchDims, cudaStream_t *stream, int opNum, void *x, Nd4jLong *xShape, int xRank, void *extraParams, void *z, Nd4jLong *zShape, int zRank, int *allocationPointer, void *reductionPointer, Nd4jLong *tadShapeInfo, Nd4jLong *tadOffsets) {
DISPATCH_BY_OPNUM_T(intermediateShaped, PARAMS(launchDims, stream, x, xShape, xRank, extraParams, z, zShape, zRank, allocationPointer, reductionPointer, tadShapeInfo, tadOffsets), TRANSFORM_SAME_OPS);
DEBUG_KERNEL(stream, opNum);
}
template<typename X>
template <typename OpType>
__device__ void TransformSame<X>::transformCuda(void *vx, Nd4jLong *xShapeInfo,
void *vparams,
void *vz, Nd4jLong *zShapeInfo,
int *allocationPointer, void *vreductionPointer,
Nd4jLong *tadShapeInfo, Nd4jLong *tadOffsets) {
auto x = static_cast<X*>(vx);
auto z = static_cast<X*>(vz);
auto params = static_cast<X*>(vparams);
auto reductionPointer = static_cast<X*>(vreductionPointer);
if(OpType::requiresSpecial) {
OpType::execSpecialCuda(x,xShapeInfo,z,zShapeInfo,params, allocationPointer, reductionPointer, tadShapeInfo, tadOffsets);
return;
} else {
__shared__ Nd4jLong xEws;
__shared__ Nd4jLong zEws;
__shared__ char xOrder;
__shared__ char zOrder;
__shared__ Nd4jLong length;
if (threadIdx.x == 0) {
xEws = shape::elementWiseStride(xShapeInfo);
zEws = shape::elementWiseStride(zShapeInfo);
xOrder = shape::order(xShapeInfo);
zOrder = shape::order(zShapeInfo);
length = shape::length(xShapeInfo);
}
__syncthreads();
auto tid = blockIdx.x * blockDim.x + threadIdx.x;
int totalThreads = gridDim.x * blockDim.x;
Oleh tenzor mmul (#231) * Libnd4j: TensorMMul backprop op #8174, raw implementation Signed-off-by: Oleg <oleg.semeniv@gmail.com> * Libnd4j: TensorMMul backprop op #8174 merge master and some corrections Signed-off-by: Oleg <oleg.semeniv@gmail.com> * Libnd4j: TensorMMul backprop op #8174 algorithm update, need testing, sync with master * Libnd4j: TensorMMul backprop op #8174 fixed incorrect B axes calculation Signed-off-by: Oleg <oleg.semeniv@gmail.com> * Libnd4j: TensorMMul backprop op #8174 optimize axes identification and fix bug of indeces overlapping, added first test. need testing with different shapes Signed-off-by: Oleg <oleg.semeniv@gmail.com> * Libnd4j: TensorMMul backprop op #8174 some fixes and improvements need more testing Signed-off-by: Oleg <oleg.semeniv@gmail.com> * Libnd4j: TensorMMul backprop op #8174 fixed order of matrix multiply Signed-off-by: Oleg <oleg.semeniv@gmail.com> * Libnd4j: TensorMMul backprop op #8174 fixed issue of incorrect axes definition, add tests based on TF, need additional testing for case dLdC not equal 1 Signed-off-by: Oleg <oleg.semeniv@gmail.com> * Libnd4j: TensorMMul backprop op #8174 fixed scalar case add test Signed-off-by: Oleg <oleg.semeniv@gmail.com> * Libnd4j: TensorMMul backprop op #8174 fixed bp algorithm, axes definition, need some mode testing with different orders combination f,c; c,f f,f and add some checks for inputs Signed-off-by: Oleg <oleg.semeniv@gmail.com> * Libnd4j: TensorMMul backprop op #8174 some checks and corrections added tests, exists the problem with different input orders support A-f B-c and A-f B-f Signed-off-by: Oleg <oleg.semeniv@gmail.com> * Libnd4j: TensorMMul backprop op #8174 sync master Signed-off-by: Oleg <oleg.semeniv@gmail.com> * - correct bug in MmulHelper::tensorDot(a, b, c, axes_a, axes_b,permutForC) Signed-off-by: Yurii <iuriish@yahoo.com> * Libnd4j: TensorMMul backprop op #8174 code clean up and refactoring Signed-off-by: Oleg <oleg.semeniv@gmail.com> * - add check for linspase ordered permutations in ShapeUtils::evalShapeForTensorDot Signed-off-by: Yurii <iuriish@yahoo.com> * - provide additional code in shape::reshape stuff in order to reduce amount of allocation/copy operations during reshaping procedure Signed-off-by: Yurii <iuriish@yahoo.com> * - further work on problem of wrong shape evaluation during permute/reshape procedures Signed-off-by: Yurii <iuriish@yahoo.com> * - still looking for bug reason in reshape/permute stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - correct bug in transform cuda native ops Signed-off-by: Yurii <iuriish@yahoo.com> * - correct bug in NDArray::assign Signed-off-by: Yurii <iuriish@yahoo.com> * - remove old shape::reshape stuff Signed-off-by: Yurii <iuriish@yahoo.com> * - add possibility to disable copy of old buffer to new buffer during reshape operation in NDArray class Signed-off-by: Yurii <iuriish@yahoo.com> * - correct bug in tensorDot which had to do with wrong pointers assigments Signed-off-by: Yurii <iuriish@yahoo.com> Co-authored-by: Oleh <oleg.semeniv@gmail.com>
2020-02-13 18:33:54 +01:00
if(xEws > 0 && zEws > 0 && xOrder == zOrder && xOrder == 'c') {
2019-06-06 14:21:15 +02:00
for (int i = tid; i < length; i += totalThreads)
z[i * zEws] = OpType::op(x[i * xEws], params);
}
else {
if(vx == vz) {
for (Nd4jLong i = tid; i < length; i+= totalThreads) {
auto xOffset = shape::getIndexOffset(i, xShapeInfo);
2019-06-06 14:21:15 +02:00
z[xOffset] = OpType::op(x[xOffset], params);
}
}
else {
for (Nd4jLong i = tid; i < length; i+= totalThreads) {
auto xOffset = shape::getIndexOffset(i, xShapeInfo);
auto zOffset = shape::getIndexOffset(i, zShapeInfo);
2019-06-06 14:21:15 +02:00
z[zOffset] = OpType::op(x[xOffset], params);
}
}
}
}
};
template<typename X>
template <typename OpType>
_CUDA_H void TransformSame<X>::intermediateShaped(dim3 launchDims, cudaStream_t *stream, void *x, Nd4jLong *xShape, int xRank, void *extraParams, void *z, Nd4jLong *zShape, int zRank, int *allocationPointer, void *reductionPointer, Nd4jLong *tadShapeInfo, Nd4jLong *tadOffsets) {
transformSameSimple<X, OpType><<<launchDims.x, launchDims.x, launchDims.z, *stream>>>(x, xShape, xRank, extraParams, z, zShape, zRank, allocationPointer, reductionPointer, tadShapeInfo, tadOffsets);
sd::DebugHelper::checkErrorCode(stream, "transformSame(...) failed");
2019-06-06 14:21:15 +02:00
}
BUILD_SINGLE_TEMPLATE(template class ND4J_EXPORT TransformSame, , LIBND4J_TYPES);
}
}