92 lines
3.0 KiB
Plaintext
92 lines
3.0 KiB
Plaintext
/* ******************************************************************************
|
|
*
|
|
*
|
|
* This program and the accompanying materials are made available under the
|
|
* terms of the Apache License, Version 2.0 which is available at
|
|
* https://www.apache.org/licenses/LICENSE-2.0.
|
|
*
|
|
* See the NOTICE file distributed with this work for additional
|
|
* information regarding copyright ownership.
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
|
|
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
|
|
* License for the specific language governing permissions and limitations
|
|
* under the License.
|
|
*
|
|
* SPDX-License-Identifier: Apache-2.0
|
|
******************************************************************************/
|
|
|
|
//
|
|
// @author raver119@gmail.com
|
|
// @author Yurii Shyrma, created on 15.11.2018
|
|
//
|
|
|
|
#include <loops/special_kernels.h>
|
|
|
|
namespace sd {
|
|
|
|
///////////////////////////////////////////////////////////////////////
|
|
/**
|
|
* This kernel accumulates X arrays, and stores z into Z
|
|
*
|
|
* @tparam T
|
|
* @param x
|
|
* @param z
|
|
* @param n
|
|
* @param length
|
|
*/
|
|
template<typename T>
|
|
__device__ void accumulateKernel(void **vx, void *vz, int n, const Nd4jLong length) {
|
|
|
|
auto x = reinterpret_cast<T **>(vx);
|
|
auto z = reinterpret_cast<T *>(vz);
|
|
|
|
__shared__
|
|
T *shmem;
|
|
|
|
if (threadIdx.x == 0) {
|
|
extern __shared__ unsigned char sharedmem[];
|
|
shmem = (T *) sharedmem;
|
|
}
|
|
__syncthreads();
|
|
|
|
for (int r = blockDim.x * blockIdx.x; r < length; r += blockDim.x * gridDim.x) {
|
|
shmem[threadIdx.x] = 0.0f;
|
|
|
|
Nd4jLong baseIdx = r;
|
|
|
|
// aggregation step, we roll over all arrays
|
|
for (int ar = 0; ar < n; ar++) {
|
|
T *cdata = (T *) x[ar];
|
|
cdata += baseIdx;
|
|
|
|
if (baseIdx + threadIdx.x < length)
|
|
shmem[threadIdx.x] += cdata[threadIdx.x];
|
|
}
|
|
|
|
T *wdata = z + baseIdx;
|
|
|
|
// saving accumulated values
|
|
if (baseIdx + threadIdx.x < length)
|
|
wdata[threadIdx.x] = shmem[threadIdx.x];
|
|
}
|
|
}
|
|
|
|
///////////////////////////////////////////////////////////////////////
|
|
template<typename T>
|
|
__global__ void execAccumulateKernel(void **vx, void *vz, int n, const Nd4jLong length) {
|
|
|
|
accumulateKernel<T>(vx, vz, n, length);
|
|
}
|
|
|
|
///////////////////////////////////////////////////////////////////////
|
|
template<typename T>
|
|
__host__ void
|
|
accumulateKernelGeneric(dim3 &launchDims, cudaStream_t *stream, void **vx, void *vz, int n, const Nd4jLong length) {
|
|
|
|
execAccumulateKernel<T><<< launchDims.x, launchDims.y, launchDims.z, *stream>>> (vx, vz, n, length);
|
|
sd::DebugHelper::checkErrorCode(stream, "accumulate(...) failed");
|
|
}
|
|
|
|
BUILD_SINGLE_TEMPLATE(template void ND4J_EXPORT accumulateKernelGeneric, (dim3 & launchDims, cudaStream_t * stream, void * *vx, void * vz, int n, const Nd4jLong length), LIBND4J_TYPES);
|
|
} |