resort .cu headers, set clang-format not sort include block and consider .cu...

resort .cu headers, set clang-format not sort include block and consider .cu as main source file (#43633)

.pre-commit-config.yaml

@@ -27,7 +27,7 @@ repos:
         files: (api|backward)\.yaml$
 -   repo: local
     hooks:
-    -   id: clang-format-with-version-check
+    -   id: clang-format
         name: clang-format
         description: Format files with ClangFormat.
         entry: bash ./tools/codestyle/clang_format.hook -i

paddle/phi/kernels/gpu/abs_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/abs_grad_kernel.h"
+
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/abs_grad_kernel.h"
 #include "paddle/phi/kernels/impl/abs_grad_kernel_impl.h"
 
 using phi::dtype::complex;

paddle/phi/kernels/gpu/abs_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/abs_kernel.h"
+
 #include <algorithm>
 #include <vector>
 
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/abs_kernel.h"
 #include "paddle/phi/kernels/funcs/complex_functors.h"
 #include "paddle/phi/kernels/funcs/elementwise_base.h"
 

paddle/phi/kernels/gpu/accuracy_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/accuracy_kernel.h"
+
 #include <thrust/execution_policy.h>
 #include <thrust/reduce.h>
 
@@ -20,7 +22,6 @@
 #include "paddle/phi/backends/gpu/gpu_info.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/accuracy_kernel.h"
 
 namespace phi {
 using paddle::platform::PADDLE_CUDA_NUM_THREADS;

paddle/phi/kernels/gpu/activation_grad_kernel.cu

@@ -12,12 +12,13 @@ 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. */
 
+#include "paddle/phi/kernels/activation_grad_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_device_function.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/activation_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_base.h"
 #include "paddle/phi/kernels/impl/activation_grad_impl.h"
 

paddle/phi/kernels/gpu/activation_kernel.cu

@@ -12,12 +12,13 @@ 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. */
 
+#include "paddle/phi/kernels/activation_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_device_function.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/activation_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_base.h"
 #include "paddle/phi/kernels/impl/activation_grad_impl.h"
 #include "paddle/phi/kernels/impl/activation_impl.h"

paddle/phi/kernels/gpu/adadelta_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/adadelta_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/adadelta_kernel.h"
 #include "paddle/phi/kernels/impl/adadelta_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/adagrad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/adagrad_kernel.h"
+
 #include "paddle/fluid/operators/math/selected_rows_functor.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/adagrad_kernel.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/impl/adagrad_kernel_impl.h"
 

paddle/phi/kernels/gpu/adam_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/adam_kernel.h"
+
 #include <math.h>  // for sqrt in CPU and CUDA
 
 #include <vector>
@@ -22,7 +24,6 @@
 #include "paddle/phi/common/amp_type_traits.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/adam_kernel.h"
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/adam_functors.h"
 #include "paddle/phi/kernels/funcs/for_range.h"

paddle/phi/kernels/gpu/adamax_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/adamax_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/adamax_kernel.h"
 #include "paddle/phi/kernels/impl/adamax_kernel_impl.h"
 
 PD_REGISTER_KERNEL(adamax, GPU, ALL_LAYOUT, phi::AdamaxKernel, float, double) {}

paddle/phi/kernels/gpu/adamw_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/adamw_kernel.h"
+
 #include <math.h>  // for sqrt in CPU and CUDA
 
 #include <vector>
@@ -22,7 +24,6 @@
 #include "paddle/phi/common/amp_type_traits.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/adamw_kernel.h"
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/adam_functors.h"
 #include "paddle/phi/kernels/funcs/for_range.h"

paddle/phi/kernels/gpu/add_n_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/add_n_kernel.h"
+
 #include "paddle/fluid/memory/malloc.h"
 #include "paddle/fluid/memory/memcpy.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/add_n_kernel.h"
 #include "paddle/phi/kernels/funcs/eigen/common.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 

paddle/phi/kernels/gpu/addmm_grad_kernel.cu

@@ -12,9 +12,10 @@ 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. */
 
+#include "paddle/phi/kernels/addmm_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/addmm_grad_kernel.h"
 #include "paddle/phi/kernels/impl/addmm_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/addmm_kernel.cu

@@ -12,9 +12,10 @@ 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. */
 
+#include "paddle/phi/kernels/addmm_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/addmm_kernel.h"
 #include "paddle/phi/kernels/impl/addmm_kernel_impl.h"
 
 PD_REGISTER_KERNEL(addmm, GPU, ALL_LAYOUT, phi::AddmmKernel, float, double) {}

paddle/phi/kernels/gpu/allclose_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/allclose_kernel.h"
+
 #include "paddle/phi/core/enforce.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/allclose_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/arange_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/arange_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/arange_kernel.h"
 #include "paddle/phi/kernels/copy_kernel.h"
 #include "paddle/phi/kernels/funcs/range_function.h"
 

paddle/phi/kernels/gpu/arg_min_max_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/arg_min_max_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/arg_min_max_kernel.h"
 
 #if defined(__NVCC__) || defined(__HIPCC__)
 

paddle/phi/kernels/gpu/argsort_kernel.cu

@@ -12,12 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/argsort_kernel.h"
+
 #include <thrust/copy.h>
 #include <thrust/execution_policy.h>
 #include <thrust/sequence.h>
 #include <thrust/sort.h>
-
-#include "paddle/phi/kernels/argsort_kernel.h"
 #ifdef __NVCC__
 #include "cub/cub.cuh"
 #endif

paddle/phi/kernels/gpu/auc_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/auc_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/cpu/cpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/auc_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/bce_loss_grad_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/bce_loss_grad_kernel.h"
+
 #include <algorithm>
 #include <vector>
 
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/hostdevice.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/bce_loss_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_base.h"
 
 namespace phi {

paddle/phi/kernels/gpu/bce_loss_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/bce_loss_kernel.h"
+
 #include <algorithm>
 #include <vector>
 
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/hostdevice.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/bce_loss_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_base.h"
 #include "paddle/phi/kernels/primitive/functor_primitives.h"
 

paddle/phi/kernels/gpu/bilinear_tensor_product_grad_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/bilinear_tensor_product_grad_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/bilinear_tensor_product_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(bilinear_tensor_product_grad,

paddle/phi/kernels/gpu/bilinear_tensor_product_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/bilinear_tensor_product_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/bilinear_tensor_product_kernel_impl.h"
 
 PD_REGISTER_KERNEL(bilinear_tensor_product,

paddle/phi/kernels/gpu/bincount_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/bincount_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/bincount_kernel.h"
 #include "paddle/phi/kernels/funcs/eigen/common.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 

paddle/phi/kernels/gpu/broadcast_tensors_grad_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/broadcast_tensors_grad_kernel.h"
+
 #include <vector>
 
 #include "paddle/fluid/framework/tensor_util.h"
@@ -19,7 +21,6 @@
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/enforce.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/broadcast_tensors_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/reduce_function.h"
 #include "paddle/phi/kernels/primitive/functor_primitives.h"
 

paddle/phi/kernels/gpu/broadcast_tensors_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/broadcast_tensors_kernel.h"
+
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/broadcast_tensors_kernel.h"
 #include "paddle/phi/kernels/impl/broadcast_tensors_kernel_impl.h"
 
 PD_REGISTER_KERNEL(broadcast_tensors,

paddle/phi/kernels/gpu/cast_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cast_grad_kernel.h"
+
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/core/visit_type.h"
-#include "paddle/phi/kernels/cast_grad_kernel.h"
 #include "paddle/phi/kernels/gpu/cast_impl.h"
 
 namespace phi {

paddle/phi/kernels/gpu/cast_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cast_kernel.h"
+
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/core/visit_type.h"
-#include "paddle/phi/kernels/cast_kernel.h"
 #include "paddle/phi/kernels/gpu/cast_impl.h"
 
 namespace phi {

paddle/phi/kernels/gpu/channel_shuffle_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/channel_shuffle_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/channel_shuffle_grad_kernel.h"
 #include "paddle/phi/kernels/impl/channel_shuffle_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(channel_shuffle_grad,

paddle/phi/kernels/gpu/channel_shuffle_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/channel_shuffle_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/channel_shuffle_kernel.h"
 #include "paddle/phi/kernels/impl/channel_shuffle_kernel_impl.h"
 
 PD_REGISTER_KERNEL(channel_shuffle,

paddle/phi/kernels/gpu/cholesky_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cholesky_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/cholesky_grad_kernel.h"
 #include "paddle/phi/kernels/impl/cholesky_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/cholesky_kernel.cu

@@ -15,6 +15,8 @@ limitations under the License. */
 #ifndef PADDLE_WITH_HIP
 // HIP not support cusolver
 
+#include "paddle/phi/kernels/cholesky_kernel.h"
+
 #include <thrust/device_vector.h>
 
 #include <algorithm>
@@ -25,7 +27,6 @@ limitations under the License. */
 #include "paddle/phi/backends/dynload/cusolver.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/cholesky_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/clip_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/clip_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/clip_grad_kernel.h"
 #include "paddle/phi/kernels/impl/clip_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(clip_grad,

paddle/phi/kernels/gpu/clip_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/clip_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/clip_kernel.h"
 #include "paddle/phi/kernels/impl/clip_kernel_impl.h"
 
 PD_REGISTER_KERNEL(clip,

paddle/phi/kernels/gpu/complex_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/complex_grad_kernel.h"
+
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/complex_grad_kernel.h"
 #include "paddle/phi/kernels/impl/complex_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(imag_grad,

paddle/phi/kernels/gpu/complex_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/complex_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/complex_kernel.h"
 #include "paddle/phi/kernels/impl/complex_kernel_impl.h"
 
 // See Note [ Why still include the fluid headers? ]

paddle/phi/kernels/gpu/concat_grad_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/concat_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/concat_grad_kernel.h"
 #include "paddle/phi/kernels/impl/concat_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(concat_grad,

paddle/phi/kernels/gpu/concat_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/concat_kernel.h"
+
 #include "paddle/fluid/operators/strided_memcpy.h"
 #include "paddle/fluid/platform/bfloat16.h"
 #include "paddle/fluid/platform/complex.h"
@@ -20,7 +22,6 @@
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/core/lod_utils.h"
-#include "paddle/phi/kernels/concat_kernel.h"
 #include "paddle/phi/kernels/funcs/concat_and_split_functor.h"
 #include "paddle/phi/kernels/funcs/concat_funcs.h"
 

paddle/phi/kernels/gpu/conv_grad_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/conv_grad_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/conv_grad_grad_kernel.h"
 #include "paddle/phi/kernels/impl/conv_grad_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/conv_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/conv_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/conv_grad_kernel.h"
 #include "paddle/phi/kernels/impl/conv_grad_kernel_impl.h"
 
 namespace phi {

paddle/phi/kernels/gpu/conv_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/conv_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/conv_kernel.h"
 #include "paddle/phi/kernels/impl/conv_kernel_impl.h"
 
 namespace phi {

paddle/phi/kernels/gpu/conv_transpose_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/conv_transpose_grad_kernel.h"
+
 #include "paddle/phi/common/layout.h"
 #include "paddle/phi/core/ddim.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/conv_transpose_grad_kernel.h"
 #include "paddle/phi/kernels/cpu/conv_util.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/gpu/depthwise_conv.h"

paddle/phi/kernels/gpu/conv_transpose_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/conv_transpose_kernel.h"
+
 #include "paddle/phi/common/layout.h"
 #include "paddle/phi/core/ddim.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/conv_transpose_kernel.h"
 #include "paddle/phi/kernels/cpu/conv_util.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/gpu/depthwise_conv.h"

paddle/phi/kernels/gpu/copy_kernel.cu

@@ -12,11 +12,12 @@ 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. */
 
+#include "paddle/phi/kernels/copy_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/data_type.h"
 #include "paddle/phi/core/compat/convert_utils.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/copy_kernel.h"
 
 // See Note [ Why still include the fluid headers? ]
 #include "paddle/fluid/memory/malloc.h"

paddle/phi/kernels/gpu/cross_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cross_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/cross_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/reduce_function.h"
 
 namespace phi {

paddle/phi/kernels/gpu/cross_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cross_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/cross_kernel.h"
 #include "paddle/phi/kernels/funcs/reduce_function.h"
 
 namespace phi {

paddle/phi/kernels/gpu/cum_kernel.cu

@@ -12,12 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cum_kernel.h"
+
 #include <thrust/device_ptr.h>
 #include <thrust/device_vector.h>
 #include <thrust/reverse.h>
 #include <thrust/scan.h>
-
-#include "paddle/phi/kernels/cum_kernel.h"
 #ifdef __NVCC__
 #include <cub/cub.cuh>
 #endif
@@ -271,7 +271,6 @@ void ScanKernel(const Context& dev_ctx,
     return;
   }
 
-
   size_t height = 1;
   size_t width = 1;
   for (size_t i = 0; i <= axis; i++) {
@@ -391,9 +390,5 @@ PD_REGISTER_KERNEL(cumsum,
                    int,
                    int64_t) {}
 
-PD_REGISTER_KERNEL(logcumsumexp,
-                   GPU,
-                   ALL_LAYOUT,
-                   phi::LogcumsumexpKernel,
-                   float,
-                   double) {}
+PD_REGISTER_KERNEL(
+    logcumsumexp, GPU, ALL_LAYOUT, phi::LogcumsumexpKernel, float, double) {}

paddle/phi/kernels/gpu/cumprod_grad_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cumprod_grad_kernel.h"
+
 #include <thrust/transform.h>
 
 #include "paddle/fluid/operators/math/inclusive_scan.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/cumprod_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/complex_functors.h"
 #include "paddle/phi/kernels/funcs/cumprod.h"
 #include "paddle/phi/kernels/funcs/elementwise_functor.h"

paddle/phi/kernels/gpu/cumprod_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/cumprod_kernel.h"
+
 #include "paddle/fluid/operators/math/inclusive_scan.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/cumprod_kernel.h"
 #include "paddle/phi/kernels/funcs/complex_functors.h"
 #include "paddle/phi/kernels/funcs/cumprod.h"
 #include "paddle/phi/kernels/funcs/elementwise_functor.h"

paddle/phi/kernels/gpu/deformable_conv_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/deformable_conv_grad_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/deformable_conv_grad_kernel.h"
 #include "paddle/phi/kernels/impl/deformable_conv_grad_kernel_impl.h"
 
 namespace phi {

paddle/phi/kernels/gpu/deformable_conv_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/deformable_conv_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/deformable_conv_kernel.h"
 #include "paddle/phi/kernels/impl/deformable_conv_kernel_impl.h"
 
 PD_REGISTER_KERNEL(deformable_conv,

paddle/phi/kernels/gpu/determinant_grad_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/determinant_grad_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/determinant_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(determinant_grad,

paddle/phi/kernels/gpu/determinant_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/determinant_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/determinant_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/diag_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/diag_kernel.h"
+
 #include <algorithm>
 #include <tuple>
 
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/diag_kernel.h"
 #include "paddle/phi/kernels/funcs/diag_functor.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 

paddle/phi/kernels/gpu/diagonal_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/diagonal_grad_kernel.h"
+
 #include "paddle/fluid/framework/tensor_util.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/diagonal_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/diagonal.h"
 
 namespace phi {

paddle/phi/kernels/gpu/diagonal_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/diagonal_kernel.h"
+
 #include "paddle/fluid/framework/tensor_util.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/diagonal_kernel.h"
 #include "paddle/phi/kernels/funcs/diagonal.h"
 
 namespace phi {

paddle/phi/kernels/gpu/digamma_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/digamma_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/digamma_grad_kernel.h"
 #include "paddle/phi/kernels/impl/digamma_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/digamma_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/digamma_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/scalar.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/digamma_kernel.h"
 #include "paddle/phi/kernels/impl/digamma_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/dist_grad_kernel.cu

@@ -12,9 +12,10 @@ 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. */
 
+#include "paddle/phi/kernels/dist_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/dist_grad_kernel.h"
 #include "paddle/phi/kernels/impl/dist_grad_kernel_impl.h"
 
 #ifdef PADDLE_WITH_HIP

paddle/phi/kernels/gpu/dist_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/dist_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/dist_kernel.h"
 #include "paddle/phi/kernels/impl/dist_kernel_impl.h"
 
 #ifdef PADDLE_WITH_HIP

paddle/phi/kernels/gpu/dot_grad_kernel.cu

@@ -12,10 +12,11 @@ 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. */
 
+#include "paddle/phi/kernels/dot_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/dot_grad_kernel.h"
 #include "paddle/phi/kernels/impl/dot_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(dot_grad,

paddle/phi/kernels/gpu/dot_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/dot_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/dot_kernel.h"
 #include "paddle/phi/kernels/funcs/eigen/common.h"
 
 // See Note [ Why still include the fluid headers? ]

paddle/phi/kernels/gpu/dropout_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/dropout_grad_kernel.h"
+
 #include "paddle/fluid/operators/dropout_impl.cu.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/dropout_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/dropout_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/dropout_kernel.h"
+
 #include "paddle/fluid/operators/dropout_impl.cu.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/dropout_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/eigh_grad_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/eigh_grad_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/complex_functors.h"
 #include "paddle/phi/kernels/impl/eigh_grad_kernel_impl.h"
 

paddle/phi/kernels/gpu/eigh_kernel.cu

@@ -15,8 +15,9 @@
 #ifndef PADDLE_WITH_HIP
 // HIP not support cusolver
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/eigh_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/complex_functors.h"
 #include "paddle/phi/kernels/funcs/values_vectors_functor.h"
 

paddle/phi/kernels/gpu/einsum_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/einsum_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/einsum_kernel.h"
 #include "paddle/phi/kernels/impl/einsum_impl.h"
 
 PD_REGISTER_KERNEL(einsum,

paddle/phi/kernels/gpu/elementwise_add_grad_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/elementwise_add_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
-#include "paddle/phi/kernels/elementwise_add_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_functor.h"
 #include "paddle/phi/kernels/gpu/elementwise_grad.h"
 #include "paddle/phi/kernels/impl/elementwise_grad_kernel_impl.h"

paddle/phi/kernels/gpu/elementwise_divide_grad_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/elementwise_divide_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/elementwise_divide_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_functor.h"
 #include "paddle/phi/kernels/gpu/elementwise_grad.h"
 #include "paddle/phi/kernels/impl/elementwise_grad_kernel_impl.h"

paddle/phi/kernels/gpu/elementwise_grad_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/elementwise_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/copy_kernel.h"
-#include "paddle/phi/kernels/elementwise_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_functor.h"
 #include "paddle/phi/kernels/gpu/elementwise_grad.h"
 #include "paddle/phi/kernels/impl/elementwise_grad_kernel_impl.h"

paddle/phi/kernels/gpu/elementwise_multiply_grad_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/elementwise_multiply_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/complex.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/elementwise_multiply_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_functor.h"
 #include "paddle/phi/kernels/gpu/elementwise_grad.h"
 #include "paddle/phi/kernels/impl/elementwise_grad_kernel_impl.h"

paddle/phi/kernels/gpu/embedding_grad_kernel.cu

@@ -12,13 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/embedding_grad_kernel.h"
+
 #include "paddle/fluid/framework/mixed_vector.h"
 #include "paddle/fluid/memory/memcpy.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/data_type.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/embedding_grad_kernel.h"
 #include "paddle/phi/kernels/funcs/eigen/common.h"
 #include "paddle/phi/kernels/funcs/embedding_util.h"
 

paddle/phi/kernels/gpu/embedding_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/embedding_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/data_type.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/embedding_kernel.h"
 #include "paddle/phi/kernels/funcs/eigen/eigen_function.h"
 #include "paddle/phi/kernels/funcs/embedding_util.h"
 

paddle/phi/kernels/gpu/erf_grad_kernel.cu

@@ -12,10 +12,11 @@ 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. */
 
+#include "paddle/phi/kernels/erf_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/erf_grad_kernel.h"
 #include "paddle/phi/kernels/impl/erf_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(erf_grad,

paddle/phi/kernels/gpu/erf_kernel.cu

@@ -12,10 +12,11 @@ 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. */
 
+#include "paddle/phi/kernels/erf_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/erf_kernel.h"
 #include "paddle/phi/kernels/impl/erf_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/erfinv_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/erfinv_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/erfinv_grad_kernel.h"
 #include "paddle/phi/kernels/impl/erfinv_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(

paddle/phi/kernels/gpu/erfinv_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/erfinv_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/erfinv_kernel.h"
 #include "paddle/phi/kernels/impl/erfinv_kernel_impl.h"
 
 PD_REGISTER_KERNEL(erfinv, GPU, ALL_LAYOUT, phi::ErfinvKernel, float, double) {}

paddle/phi/kernels/gpu/expand_as_grad_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/expand_as_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/expand_as_grad_kernel.h"
 #include "paddle/phi/kernels/impl/expand_as_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(expand_as_grad,

paddle/phi/kernels/gpu/expand_as_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/expand_as_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/expand_as_kernel.h"
 #include "paddle/phi/kernels/impl/expand_as_kernel_impl.h"
 
 PD_REGISTER_KERNEL(expand_as,

paddle/phi/kernels/gpu/expand_grad_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/expand_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/expand_grad_kernel.h"
 #include "paddle/phi/kernels/impl/expand_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(expand_grad,

paddle/phi/kernels/gpu/expand_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/expand_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/scalar.h"
 #include "paddle/phi/core/dense_tensor.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/expand_kernel.h"
 #include "paddle/phi/kernels/impl/expand_kernel_impl.h"
 
 PD_REGISTER_KERNEL(expand,

paddle/phi/kernels/gpu/eye_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/eye_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/eye_kernel.h"
 #include "paddle/phi/kernels/impl/eye_kernel_impl.h"
 
 PD_REGISTER_KERNEL(eye,

paddle/phi/kernels/gpu/flip_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/flip_kernel.h"
+
 #include "paddle/fluid/memory/malloc.h"
 #include "paddle/fluid/memory/memcpy.h"
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/place.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/flip_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/frobenius_norm_grad_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/frobenius_norm_grad_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/frobenius_norm_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(frobenius_norm_grad,

paddle/phi/kernels/gpu/frobenius_norm_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/frobenius_norm_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/activation_functor.h"
 #include "paddle/phi/kernels/gpu/reduce.h"
 namespace phi {

paddle/phi/kernels/gpu/full_kernel.cu

@@ -12,9 +12,10 @@ 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. */
 
+#include "paddle/phi/kernels/full_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/full_kernel.h"
 #include "paddle/phi/kernels/funcs/elementwise_base.h"
 namespace phi {
 

paddle/phi/kernels/gpu/gather_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/gather_kernel.h"
+
 #include "paddle/phi/common/bfloat16.h"
 #include "paddle/phi/common/float16.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/gather.cu.h"
-#include "paddle/phi/kernels/gather_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/gather_nd_grad_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/gather_nd_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/eigen/common.h"
 #include "paddle/phi/kernels/funcs/scatter.cu.h"
-#include "paddle/phi/kernels/gather_nd_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/gather_nd_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/gather_nd_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/gather.cu.h"
 #include "paddle/phi/kernels/funcs/scatter.cu.h"
-#include "paddle/phi/kernels/gather_nd_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/gather_tree_kernel.cu

@@ -12,10 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/gather_tree_kernel.h"
+
 #include <algorithm>
 
 #include "paddle/phi/core/kernel_registry.h"
-#include "paddle/phi/kernels/gather_tree_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/gaussian_random_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/gaussian_random_kernel.h"
+
 #include <thrust/random.h>
 
 #include "paddle/fluid/framework/generator.h"
@@ -21,7 +23,6 @@
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/distribution_helper.h"
 #include "paddle/phi/kernels/funcs/index_impl.cu.h"
-#include "paddle/phi/kernels/gaussian_random_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/graph_reindex_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/graph_reindex_kernel.h"
+
 #include <thrust/copy.h>
 #include <thrust/device_vector.h>
 #include <thrust/reduce.h>
@@ -21,7 +23,6 @@
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpu/graph_reindex_funcs.h"
-#include "paddle/phi/kernels/graph_reindex_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/graph_send_recv_grad_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/graph_send_recv_grad_kernel.h"
+
 #include <algorithm>
 #include <vector>
 
@@ -19,7 +21,6 @@
 #include "paddle/phi/core/hostdevice.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpu/graph_send_recv_funcs.h"
-#include "paddle/phi/kernels/graph_send_recv_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/graph_send_recv_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/graph_send_recv_kernel.h"
+
 #include <thrust/device_vector.h>
 #include <thrust/fill.h>
 
@@ -22,7 +24,6 @@
 #include "paddle/phi/core/hostdevice.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpu/graph_send_recv_funcs.h"
-#include "paddle/phi/kernels/graph_send_recv_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/grid_sample_grad_kernel.cu

@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/grid_sample_grad_kernel.h"
+
 #include "paddle/fluid/platform/device/gpu/gpu_device_function.h"
 #include "paddle/fluid/platform/device/gpu/gpu_primitives.h"
 #include "paddle/phi/backends/gpu/gpu_info.h"
@@ -19,7 +21,6 @@
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/gpu/grid_sample_utils.h"
-#include "paddle/phi/kernels/grid_sample_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/grid_sample_kernel.cu

@@ -12,11 +12,12 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/grid_sample_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_info.h"
 #include "paddle/phi/backends/gpu/gpu_launch_config.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gpu/grid_sample_utils.h"
-#include "paddle/phi/kernels/grid_sample_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/group_norm_grad_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/group_norm_grad_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/layout.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/gpu/group_norm_utils.h"
-#include "paddle/phi/kernels/group_norm_grad_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/group_norm_kernel.cu

@@ -12,12 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/group_norm_kernel.h"
+
 #include "paddle/phi/backends/gpu/gpu_context.h"
 #include "paddle/phi/common/layout.h"
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/math_function.h"
 #include "paddle/phi/kernels/gpu/group_norm_utils.h"
-#include "paddle/phi/kernels/group_norm_kernel.h"
 
 namespace phi {
 

paddle/phi/kernels/gpu/gumbel_softmax_grad_kernel.cu

@@ -12,8 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/gumbel_softmax_grad_kernel.h"
+
+#include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/impl/gumbel_softmax_grad_kernel_impl.h"
 
 PD_REGISTER_KERNEL(gumbel_softmax_grad,

paddle/phi/kernels/gpu/gumbel_softmax_kernel.cu

@@ -12,9 +12,10 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#include "paddle/phi/kernels/gumbel_softmax_kernel.h"
+
 #include "paddle/phi/core/kernel_registry.h"
 #include "paddle/phi/kernels/funcs/axis_utils.h"
-#include "paddle/phi/kernels/gumbel_softmax_kernel.h"
 #include "paddle/phi/kernels/impl/gumbel_softmax_kernel_impl.h"
 
 #if defined(__NVCC__) || defined(__HIPCC__)