Skip to content

Optimize GroupNormOp#12844

Closed
xiaomengy wants to merge 1 commit intopytorch:masterfrom
xiaomengy:export-D10455567
Closed

Optimize GroupNormOp#12844
xiaomengy wants to merge 1 commit intopytorch:masterfrom
xiaomengy:export-D10455567

Conversation

@xiaomengy
Copy link
Copy Markdown
Contributor

Summary: Optimize GroupNormOp

Differential Revision: D10455567

@xiaomengy
Copy link
Copy Markdown
Contributor Author

Some benchmark

On CPU, N = 128, C = 128, H = 64, W = 64, group = 32:
NCHW 214.66ms -> 152.50ms
NHWC 912.79ms -> 848.54ms

On GPU, N = 256, C = 256, H = 128, W = 128, group = 32:
NCHW 134.58ms -> 70.74ms
NHWC 221.92ms -> 102.61ms

Copy link
Copy Markdown
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.

LG, the failures are not related. Great improvement.

Summary:
Pull Request resolved: pytorch#12844

Optimize GroupNormOp

Reviewed By: houseroad

Differential Revision: D10455567

fbshipit-source-id: 40d075fb1e2fc01bc4c2a32e871a38c1ff9c91ef
@xiaomengy xiaomengy deleted the export-D10455567 branch October 19, 2018 20:57
@ezyang ezyang added the merged label Jun 25, 2019
laurentdupin pushed a commit to laurentdupin/pytorch that referenced this pull request Apr 24, 2026
Summary:
Pull Request resolved: pytorch#12844

Optimize GroupNormOp

Reviewed By: houseroad

Differential Revision: D10455567

fbshipit-source-id: aee211badd1e0c8ea6196843e3e77f7c612a74d5
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.

3 participants