Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[cherry-pick] set device id of Place() to get GPUContext needed by LimitGridDim in ElemwiseGradBroadcast #42332

Merged
merged 1 commit into from
Apr 28, 2022
Merged

[cherry-pick] set device id of Place() to get GPUContext needed by LimitGridDim in ElemwiseGradBroadcast #42332

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 3 additions & 2 deletions paddle/phi/kernels/funcs/elementwise_grad_base.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,6 +15,7 @@ limitations under the License. */
#pragma once

#include "paddle/phi/backends/all_context.h"
#include "paddle/phi/backends/gpu/gpu_info.h"
#include "paddle/phi/core/dense_tensor.h"
#include "paddle/phi/kernels/funcs/common_shape.h"
#include "paddle/phi/kernels/funcs/elementwise_utils.h"
Expand Down Expand Up @@ -978,7 +979,7 @@ static void ElemwiseGradBroadcast1CUDA(gpuStream_t stream,
// suppose perfoemance improves with h increased.
dim3 block_size = dim3(BLOCK_X, BLOCK_Y);
dim3 grid_size = dim3((w + BLOCK_X - 1) / BLOCK_X);
auto gplace = phi::GPUPlace();
auto gplace = phi::GPUPlace(phi::backends::gpu::GetCurrentDeviceId());
auto *ctx = static_cast<GPUContext *>(
paddle::platform::DeviceContextPool::Instance().Get(gplace));
paddle::platform::LimitGridDim(*ctx, &grid_size);
Expand All @@ -1003,7 +1004,7 @@ static void ElemwiseGradBroadcast2CUDA(gpuStream_t stream,
T *dy) {
int block_size = std::min(ELEMWISE_MAX_BLOCK_DIM, pre * post);
dim3 grid_size = dim3(n);
auto gplace = phi::GPUPlace();
auto gplace = phi::GPUPlace(phi::backends::gpu::GetCurrentDeviceId());
auto *ctx = static_cast<GPUContext *>(
paddle::platform::DeviceContextPool::Instance().Get(gplace));
paddle::platform::LimitGridDim(*ctx, &grid_size);
Expand Down