Optimize where_op and abs_grad_op by the elementwise interface #39609
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters. Learn more about bidirectional Unicode characters
| Thanks for your contribution! |
huangxu96 force-pushed the drop_out_kp branch from 6564ee3 to f376b92 Compare huangxu96 changed the title huangxu96 force-pushed the drop_out_kp branch from f376b92 to 0a79619 Compare 
AnnaTrainingG previously approved these changes Feb 23, 2022
| #include "paddle/phi/kernels/abs_grad_kernel.h" | ||
| #include "paddle/phi/kernels/funcs/complex_functors.h" | ||
| #if defined(__NVCC__) || defined(__HIPCC__) | ||
| #include "paddle/fluid/operators/elementwise/elementwise_op_impl.cu.h" |
Contributor
There was a problem hiding this comment.
phi下不能include fluid路径下的文件，参考cast 修改一下
Contributor
There was a problem hiding this comment.
PR描述里面介绍清楚一点做的工作，比如：添加哪些functor，调用哪个Kernel等
| auto functor = CondFunctor<T>(); | ||
| std::vector<const framework::Tensor*> ins = {condition, X, Y}; | ||
| std::vector<framework::Tensor*> outs = {out}; | ||
| paddle::operators::LaunchSameDimsElementwiseCudaKernel<T>(dev_ctx, ins, |
AnnaTrainingG approved these changes Feb 24, 2022
| T><<<config.block_per_grid.x, config.thread_per_block.x, 0, stream>>>( | ||
| numel, cond_data, x_data, y_data, out_data); | ||
| auto functor = CondFunctor<T>(); | ||
| std::vector<const framework::Tensor*> ins = {condition, X, Y}; |
Member
There was a problem hiding this comment.
相关的framework Tensor后续可以改成DensorTensor
Liu-xiandong approved these changes Feb 24, 2022
xingfeng01 approved these changes Feb 24, 2022
| | ||
| namespace paddle { | ||
| namespace operators { | ||
| |
Xreki reviewed Feb 24, 2022
| }; | ||
| | ||
| template <typename T> | ||
| __global__ void WhereCUDAKernel(const int N, const bool* cond, const T* x, |
| | ||
| template <typename T> | ||
| struct CondFunctor { | ||
| HOSTDEVICE inline CondFunctor() {} |
| | ||
| template <typename T> | ||
| struct AbsGradCUDAFunctor { | ||
| HOSTDEVICE inline AbsGradCUDAFunctor() {} |
| }; | ||
| | ||
| template <> | ||
| struct AbsGradCUDAFunctor<phi::dtype::complex<float>> { |
Contributor
There was a problem hiding this comment.
Functor定义可以简化下，参考：
Paddle/paddle/phi/kernels/gpu/abs_kernel.cu
Lines 29 to 34 in bbe441f
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
5 participants
Add this suggestion to a batch that can be applied as a single commit. This suggestion is invalid because no changes were made to the code. Suggestions cannot be applied while the pull request is closed. Suggestions cannot be applied while viewing a subset of changes. Only one suggestion per line can be applied in a batch. Add this suggestion to a batch that can be applied as a single commit. Applying suggestions on deleted lines is not supported. You must change the existing code in this line in order to create a valid suggestion. Outdated suggestions cannot be applied. This suggestion has been applied or marked resolved. Suggestions cannot be applied from pending reviews. Suggestions cannot be applied on multi-line comments. Suggestions cannot be applied while the pull request is queued to merge. Suggestion cannot be applied right now. Please check back later.
PR types
Performance optimization
PR changes
OPs
Describe
通过elementwise 接口优化了wehere_op和abs_grad_op。 elementwise 接口打包了一系列性能优化技巧，对于有elementwise行为的op有通用的性能提升。通过重写functor的形式，将代码里的循环遍历元素改写为通过elementwise接口调用functor实现。