Skip to content
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

Jump to bottom

[prototype] Minor change on adjust_saturation_image_tensor uint8 #6940

Merged
merged 2 commits into from
Nov 10, 2022
Merged

[prototype] Minor change on adjust_saturation_image_tensor uint8 #6940

merged 2 commits into from
Nov 10, 2022

Conversation

datumbox
Copy link
Contributor

@datumbox datumbox commented Nov 10, 2022 •
edited by pytorch-bot bot
Loading

Related to #6818

Minor perf/code-quality improvement for adjust_saturation_image_tensor on uint8. I should have done it as part of #6933 which introduced the parameter but missed it. Floats remain unaffected:

[---------------- adjust_saturation_image_tensor cpu torch.float32 ---------------]
                         |  adjust_saturation_image_tensor old  |      fn2 new     
1 threads: ------------------------------------------------------------------------
      (16, 3, 400, 400)  |           14243 (+-425) us           |  14257 (+-265) us
      (3, 400, 400)      |            569 (+-  1) us            |   567 (+-  1) us 
6 threads: ------------------------------------------------------------------------
      (16, 3, 400, 400)  |           14924 (+-250) us           |  14919 (+-308) us
      (3, 400, 400)      |            805 (+- 11) us            |   802 (+- 10) us 

Times are in microseconds (us).

[-------------- adjust_saturation_image_tensor cuda torch.float32 --------------]
                         |  adjust_saturation_image_tensor old  |     fn2 new    
1 threads: ----------------------------------------------------------------------
      (16, 3, 400, 400)  |            198 (+-  0) us            |  198 (+-  0) us
      (3, 400, 400)      |             54 (+-  1) us            |   53 (+-  1) us
6 threads: ----------------------------------------------------------------------
      (16, 3, 400, 400)  |            198 (+-  0) us            |  198 (+-  0) us
      (3, 400, 400)      |             54 (+-  1) us            |   53 (+-  1) us

Times are in microseconds (us).

[---------------- adjust_saturation_image_tensor cpu torch.uint8 ---------------]
                         |  adjust_saturation_image_tensor old  |     fn2 new    
1 threads: ----------------------------------------------------------------------
      (16, 3, 400, 400)  |             28 (+-  0) ms            |   27 (+-  0) ms
      (3, 400, 400)      |              1 (+-  0) ms            |    1 (+-  0) ms
6 threads: ----------------------------------------------------------------------
      (16, 3, 400, 400)  |             32 (+-  0) ms            |   30 (+-  0) ms
      (3, 400, 400)      |              2 (+-  0) ms            |    2 (+-  0) ms

Times are in milliseconds (ms).

[--------------- adjust_saturation_image_tensor cuda torch.uint8 ---------------]
                         |  adjust_saturation_image_tensor old  |     fn2 new    
1 threads: ----------------------------------------------------------------------
      (16, 3, 400, 400)  |            245 (+-  0) us            |  241 (+-  0) us
      (3, 400, 400)      |             72 (+-  1) us            |   70 (+-  1) us
6 threads: ----------------------------------------------------------------------
      (16, 3, 400, 400)  |            246 (+-  0) us            |  241 (+-  0) us
      (3, 400, 400)      |             72 (+-  1) us            |   69 (+-  1) us

Times are in microseconds (us).

cc @vfdev-5 @bjuncek @pmeier

@datumbox datumbox requested a review from pmeier November 10, 2022 11:31
@datumbox datumbox changed the title Minor change on adjust_saturation_image_tensor uint8 [prototype] Minor change on adjust_saturation_image_tensor uint8 Nov 10, 2022
@datumbox datumbox mentioned this pull request Nov 10, 2022
Closed
31 tasks
@datumbox datumbox requested a review from vfdev-5 November 10, 2022 11:35
pmeier
pmeier approved these changes Nov 10, 2022
View reviewed changes
Copy link
Collaborator

@pmeier pmeier left a comment

Choose a reason for hiding this comment

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

Thanks Vasilis!

@datumbox datumbox merged commit 0cc9080 into pytorch:main Nov 10, 2022
@datumbox datumbox deleted the perf/saturation branch November 10, 2022 13:42
facebook-github-bot pushed a commit that referenced this pull request Nov 14, 2022
@facebook-github-bot
[fbsync] Minor perf on saturation uint8 (#6940)
3306159 
Reviewed By: NicolasHug

Differential Revision: D41265192

fbshipit-source-id: 2f9b157934d97790082fac56bd300dc7be8f30cf
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@pmeier pmeier pmeier approved these changes

@vfdev-5 vfdev-5 Awaiting requested review from vfdev-5

Assignees
No one assigned
Labels
cla signed code quality module: transforms prototype
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@datumbox @pmeier @facebook-github-bot