Skip to content

【CUDA Kernel No.56、57、66、74、75】算子Kernel修复 #2191

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

Open
Le-soleile wants to merge 16 commits into
base: develop
Choose a base branch
from
Open

【CUDA Kernel No.56、57、66、74、75】算子Kernel修复 #2191

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
16 commits
Select commit Hold shift + click to select a range
e2d48b4
fix
Le-soleile Nov 19, 2025
e324cb5
Merge branch 'global_gather_kernel' of https://github.com/Le-soleile/…
Le-soleile Nov 19, 2025
ecd09b8
Merge branch 'develop' into global_gather_kernel
Le-soleile Nov 20, 2025
97cacdd
Merge branch 'develop' into global_gather_kernel
Le-soleile Nov 24, 2025
24965a6
修改路径错误
Le-soleile Nov 29, 2025
0ff3221
Merge branch 'global_gather_kernel' of https://github.com/Le-soleile/…
Le-soleile Nov 29, 2025
f06133d
Merge branch 'develop' into global_gather_kernel
Le-soleile Nov 29, 2025
eaa78c2
Merge branch 'develop' into global_gather_kernel
Le-soleile Dec 3, 2025
8db55d6
Merge branch 'develop' into global_gather_kernel
Le-soleile Dec 4, 2025
fc9bb36
Merge branch 'develop' into global_gather_kernel
YqGe585 Dec 4, 2025
961363d
Merge branch 'develop' into global_gather_kernel
YqGe585 Dec 4, 2025
fa9ca73
fix
Le-soleile Dec 5, 2025
83ffb62
Merge branch 'global_gather_kernel' of https://github.com/Le-soleile/…
Le-soleile Dec 5, 2025
fa8c1b6
Merge branch 'develop' into global_gather_kernel
Le-soleile Dec 5, 2025
ffd9d14
Merge branch 'develop' into global_gather_kernel
Le-soleile Dec 8, 2025
6f8b76f
Update CMakeLists.txt
Le-soleile Dec 8, 2025
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
2 changes: 1 addition & 1 deletion Paddle
Open in desktop
Submodule Paddle updated 976 files
2 changes: 1 addition & 1 deletion backends/iluvatar_gpu/kernels/cuda_kernels/global_gather_kernel_register.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@
// limitations under the License.

#include "paddle/phi/core/kernel_registry.h"
#include "paddle/phi/kernels/gpu/global_gather_kernel.cu" //NOLINT
#include "paddle/phi/kernels/gpu/global_gather_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(global_gather,
iluvatar_gpu,
Expand Down
2 changes: 1 addition & 1 deletion backends/iluvatar_gpu/kernels/cuda_kernels/global_scatter_kernel_register.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@
// limitations under the License.

#include "paddle/phi/core/kernel_registry.h"
#include "paddle/phi/kernels/gpu/global_scatter_kernel.cu" //NOLINT
#include "paddle/phi/kernels/gpu/global_scatter_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(global_scatter,
iluvatar_gpu,
Expand Down
2 changes: 1 addition & 1 deletion backends/iluvatar_gpu/kernels/cuda_kernels/l1_norm_grad_kernel_register.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@
// limitations under the License.

#include "paddle/phi/core/kernel_registry.h"
#include "paddle/phi/kernels/gpu/l1_norm_grad_kernel.cu" // NOLINT
#include "paddle/phi/kernels/l1_norm_grad_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(
l1_norm_grad, iluvatar_gpu, ALL_LAYOUT, phi::L1NormGradKernel, float) {}
2 changes: 1 addition & 1 deletion backends/iluvatar_gpu/kernels/cuda_kernels/lookup_table_grad_kernel_register.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@
// limitations under the License.

#include "paddle/phi/core/kernel_registry.h"
#include "paddle/phi/kernels/gpu/lookup_table_grad_kernel.cu" //NOLINT
#include "paddle/phi/kernels/gpu/lookup_table_grad_kernel.h"
PD_CUSTOM_KERNEL_REGISTER(lookup_table_grad,
iluvatar_gpu,
ALL_LAYOUT,
Expand Down
2 changes: 1 addition & 1 deletion backends/iluvatar_gpu/kernels/cuda_kernels/lookup_table_kernel_register.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@
// limitations under the License.

#include "paddle/phi/core/kernel_registry.h"
#include "paddle/phi/kernels/gpu/lookup_table_kernel.cu" //NOLINT
#include "paddle/phi/kernels/gpu/lookup_table_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(lookup_table,
iluvatar_gpu,
Expand Down
19 changes: 5 additions & 14 deletions backends/metax_gpu/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -124,20 +124,11 @@ file(
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/funcs/math/*.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/funcs/eigen/*.cu
# kernels/gpu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/partial_send_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/partial_recv_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/partial_concat_grad_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/partial_allgather_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/ctc_align_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/yolo_box_head_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/stft_grad_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/soft_relu_grad_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/shuffle_channel_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/lgamma_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/lamb_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/label_smooth_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/label_smooth_grad_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/l1_norm_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/lookup_table_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/lookup_table_grad_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/l1_norm_grad_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/global_scatter_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/global_gather_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/yolo_box_post_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/spectral_norm_grad_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/spectral_norm_kernel.cu
Expand Down
2 changes: 1 addition & 1 deletion backends/metax_gpu/kernels/cuda_kernels/global_gather_kernel_register.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@
// limitations under the License.

#include "paddle/phi/core/kernel_registry.h"
#include "paddle/phi/kernels/gpu/global_gather_kernel.cu" //NOLINT
#include "paddle/phi/kernels/gpu/global_gather_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(global_gather,
metax_gpu,
Expand Down
2 changes: 1 addition & 1 deletion backends/metax_gpu/kernels/cuda_kernels/global_scatter_kernel_register.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@
// limitations under the License.

#include "paddle/phi/core/kernel_registry.h"
#include "paddle/phi/kernels/gpu/global_scatter_kernel.cu" //NOLINT
#include "paddle/phi/kernels/gpu/global_scatter_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(global_scatter,
metax_gpu,
Expand Down
2 changes: 1 addition & 1 deletion backends/metax_gpu/kernels/cuda_kernels/l1_norm_grad_kernel_register.cu
View file
Open in desktop
Copy link
Member

@YqGe585 YqGe585 Dec 5, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

需要在metax的cmakelists中编译这个kernel，否则会导致报错：https://github.com/PaddlePaddle/PaddleCustomDevice/actions/runs/19930245237/job/57141413230?pr=2191#step:5:613

Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@
// limitations under the License.

#include "paddle/phi/core/kernel_registry.h"
#include "paddle/phi/kernels/gpu/l1_norm_grad_kernel.cu" // NOLINT
#include "paddle/phi/kernels/l1_norm_grad_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(
l1_norm_grad, metax_gpu, ALL_LAYOUT, phi::L1NormGradKernel, float) {}
2 changes: 1 addition & 1 deletion backends/metax_gpu/kernels/cuda_kernels/lookup_table_grad_kernel_register.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@
// limitations under the License.

#include "paddle/phi/core/kernel_registry.h"
#include "paddle/phi/kernels/gpu/lookup_table_grad_kernel.cu" //NOLINT
#include "paddle/phi/kernels/gpu/lookup_table_grad_kernel.h"
PD_CUSTOM_KERNEL_REGISTER(lookup_table_grad,
metax_gpu,
ALL_LAYOUT,
Expand Down
2 changes: 1 addition & 1 deletion backends/metax_gpu/kernels/cuda_kernels/lookup_table_kernel_register.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@
// limitations under the License.

#include "paddle/phi/core/kernel_registry.h"
#include "paddle/phi/kernels/gpu/lookup_table_kernel.cu" //NOLINT
#include "paddle/phi/kernels/gpu/lookup_table_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(lookup_table,
metax_gpu,
Expand Down
Loading