Skip to content

Optimize reduce ops for 2d and 3d#9992

Closed
xiaomengy wants to merge 1 commit intopytorch:masterfrom
xiaomengy:export-D9042505
Closed

Optimize reduce ops for 2d and 3d#9992
xiaomengy wants to merge 1 commit intopytorch:masterfrom
xiaomengy:export-D9042505

Conversation

@xiaomengy
Copy link
Contributor

Summary: Optimize reduce ops for 2d and 3d

Differential Revision: D9042505

@houseroad
Copy link
Member

Seems hip still has compilation errors, and incurs segmentation fault on caffe2/python/convnet_benchmarks_test.py::TestConvnetBenchmarks::testConvnetBenchmarks

@bddppq
Copy link
Contributor

bddppq commented Jul 30, 2018

@petrex @ashishfarmer @rohithkrn We probably should also hipify math_hip, otherwise we can never catch up the speed of @BIT-silence

@petrex
Copy link
Contributor

petrex commented Jul 30, 2018

thanks, we will look into this PR

@xiaomengy
Copy link
Contributor Author

I will update this PR to make build and test passed. But I cannot guarantee they can be optimized as well as the CUDA side.

@xiaomengy
Copy link
Contributor Author

@pytorchbot test this please

Summary:
Pull Request resolved: pytorch#9992

Optimize reduce ops for 2d and 3d

Differential Revision: D9042505

fbshipit-source-id: 43cbc1a8aab0d33a90a369752d2617cc564957c7
Copy link
Member

@houseroad houseroad left a comment

Choose a reason for hiding this comment

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

LGTM, thanks

@xiaomengy xiaomengy deleted the export-D9042505 branch August 4, 2018 20:57
goodlux pushed a commit to goodlux/pytorch that referenced this pull request Aug 15, 2018
Summary:
Pull Request resolved: pytorch#9992

Optimize reduce ops for 2d and 3d

Reviewed By: houseroad

Differential Revision: D9042505

fbshipit-source-id: 62af2125aa6439106293e59bdf6a2b920792fd2d
@ezyang ezyang added the merged label Jun 26, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants