Skip to content

【CUDA Kernel No.95、96、97、98、99】算子Kernel修复 #2252

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 12 commits into
base: develop
Choose a base branch
from
Open

【CUDA Kernel No.95、96、97、98、99】算子Kernel修复 #2252

Show file tree
Hide file tree
Changes from 4 commits
Commits
Show all changes
12 commits
Select commit Hold shift + click to select a range
470f8bb
fix
Le-soleile Dec 3, 2025
4253aa9
Merge branch 'develop' into 94-
Le-soleile Dec 3, 2025
56a5861
Merge branch 'develop' into 94-
YqGe585 Dec 4, 2025
1d8e84b
Merge branch 'develop' into 94-
YqGe585 Dec 4, 2025
4e6e3c3
Update quantize_linear_kernel_register.cu
Le-soleile Dec 8, 2025
4a9016c
Include quantize_linear_kernel in CUDA kernel register
Le-soleile Dec 8, 2025
e4bb89a
Merge branch 'develop' into 94-
Le-soleile Dec 8, 2025
ad557a0
Fix include statement for quantize_linear_kernel.cu
Le-soleile Dec 8, 2025
c76c8ba
Update reduce_kernel_register.cu
Le-soleile Dec 14, 2025
b70c926
Update reduce_kernel_register.cc
Le-soleile Dec 14, 2025
d8d7cdd
Merge branch 'develop' into 94-
Le-soleile Dec 14, 2025
f79c8f2
Merge branch 'develop' into 94-
YqGe585 Dec 15, 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
1 change: 0 additions & 1 deletion backends/iluvatar_gpu/kernels/cuda_kernels/quantize_linear_kernel_register.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,6 @@
// limitations under the License.

#include "paddle/phi/core/kernel_registry.h"
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.

目前有编译报错，原因是paddle/phi/kernels/quantize_linear_kernel.h头文件中声明不全，需要补充。同时需要去掉#include "paddle/phi/kernels/gpu/quantize_linear_kernel.cu"和#include "paddle/phi/kernels/impl/quantize_linear_impl.h"，并且需要确保paddle/phi/kernels/gpu/quantize_linear_kernel.cu这个文件在cmakelists中已经编译。

Copy link
Contributor Author

@Le-soleile Le-soleile Dec 8, 2025

Choose a reason for hiding this comment

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

我看有人提交了对quantize_linear_kernel的paddle仓和本仓库的修改，ci也跑完了，我这边把对quantize_linear_kernel的修改删掉了

Copy link
Contributor Author

@Le-soleile Le-soleile Dec 8, 2025 •
edited
Loading

Choose a reason for hiding this comment

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

reduce_kernel_register在paddle仓有缺少的声明，在这个pr里面PaddlePaddle/Paddle#76781 ，我就不重复提交了

#include "paddle/phi/kernels/gpu/quantize_linear_kernel.cu" //NOLINT
#include "paddle/phi/kernels/impl/quantize_linear_impl.h"
#include "paddle/phi/kernels/quantize_linear_kernel.h"

Expand Down
2 changes: 1 addition & 1 deletion backends/iluvatar_gpu/kernels/cuda_kernels/reduce_kernel_register.cc
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/reduce_kernel.cu" // NOLINT
#include "paddle/phi/kernels/reduce_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(reduce,
iluvatar_gpu,
Expand Down
2 changes: 1 addition & 1 deletion backends/iluvatar_gpu/kernels/cuda_kernels/repeat_interleave_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/repeat_interleave_grad_kernel.cu" // NOLINT
#include "paddle/phi/kernels/repeat_interleave_grad_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(repeat_interleave_with_tensor_index_grad,
iluvatar_gpu,
Expand Down
2 changes: 1 addition & 1 deletion backends/iluvatar_gpu/kernels/cuda_kernels/repeat_interleave_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/repeat_interleave_kernel.cu" // NOLINT
#include "paddle/phi/kernels/repeat_interleave_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(repeat_interleave,
iluvatar_gpu,
Expand Down
2 changes: 1 addition & 1 deletion backends/iluvatar_gpu/kernels/cuda_kernels/rmsprop_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/rmsprop_kernel.cu" // NOLINT
#include "paddle/phi/kernels/rmsprop_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(rmsprop,
iluvatar_gpu,
Expand Down
2 changes: 1 addition & 1 deletion backends/iluvatar_gpu/kernels/cuda_kernels/roi_align_grad_kernel_register.cc
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/roi_align_grad_kernel.cu" // NOLINT
#include "paddle/phi/kernels/roi_align_grad_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(
roi_align_grad, iluvatar_gpu, ALL_LAYOUT, phi::RoiAlignGradKernel, float) {}
4 changes: 4 additions & 0 deletions backends/metax_gpu/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -124,6 +124,10 @@ 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/rmsprop_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/repeat_interleave_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/repeat_interleave_grad_kernel.cu
${PADDLE_SOURCE_DIR}/paddle/phi/kernels/gpu/quantize_linear_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
Expand Down
1 change: 0 additions & 1 deletion backends/metax_gpu/kernels/cuda_kernels/quantize_linear_kernel_register.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,6 @@
// limitations under the License.

#include "paddle/phi/core/kernel_registry.h"
#include "paddle/phi/kernels/gpu/quantize_linear_kernel.cu" //NOLINT
#include "paddle/phi/kernels/impl/quantize_linear_impl.h"
#include "paddle/phi/kernels/quantize_linear_kernel.h"

Expand Down
1 change: 0 additions & 1 deletion backends/metax_gpu/kernels/cuda_kernels/reduce_kernel_register.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,7 +14,6 @@
// limitations under the License.

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

PD_CUSTOM_KERNEL_REGISTER(reduce,
Expand Down
2 changes: 1 addition & 1 deletion backends/metax_gpu/kernels/cuda_kernels/repeat_interleave_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/repeat_interleave_grad_kernel.cu" // NOLINT
#include "paddle/phi/kernels/repeat_interleave_grad_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(repeat_interleave_with_tensor_index_grad,
metax_gpu,
Expand Down
2 changes: 1 addition & 1 deletion backends/metax_gpu/kernels/cuda_kernels/repeat_interleave_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/repeat_interleave_kernel.cu" // NOLINT
#include "paddle/phi/kernels/repeat_interleave_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(repeat_interleave,
metax_gpu,
Expand Down
2 changes: 1 addition & 1 deletion backends/metax_gpu/kernels/cuda_kernels/rmsprop_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/rmsprop_kernel.cu" // NOLINT
#include "paddle/phi/kernels/rmsprop_kernel.h"

PD_CUSTOM_KERNEL_REGISTER(rmsprop,
metax_gpu,
Expand Down
Loading