Skip to content

Try using latest cu(DNN) binaries #1836

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
Merged
merged 1 commit into from
Jan 15, 2022
Merged

Try using latest cu(DNN) binaries #1836

merged 1 commit into from
Jan 15, 2022

Conversation

@ToucheSir
Copy link
Member

@ToucheSir ToucheSir commented Jan 15, 2022

Possible alternative to #1822.

@ToucheSir
Copy link
Member Author

ToucheSir commented Jan 15, 2022

bors try

bors bot added a commit that referenced this pull request Jan 15, 2022
@bors
Try #1836:
3abd317
@bors
Copy link
Contributor

bors bot commented Jan 15, 2022

try

Build succeeded:

@ToucheSir
Copy link
Member Author

ToucheSir commented Jan 15, 2022

bors r+

bors bot added a commit that referenced this pull request Jan 15, 2022
@bors @ToucheSir
Merge #1836
5108ab2 
1836: Try using latest cu(DNN) binaries r=ToucheSir a=ToucheSir

Possible alternative to #1822.

Co-authored-by: Brian Chen <ToucheSir@users.noreply.github.com>
@DhairyaLGandhi
Copy link
Member

DhairyaLGandhi commented Jan 15, 2022 •
edited
Loading

Was there some background to this? Is there something about the cudnn we were using causing this?

@DhairyaLGandhi DhairyaLGandhi marked this pull request as ready for review January 15, 2022 19:54
@DhairyaLGandhi
Copy link
Member

DhairyaLGandhi commented Jan 15, 2022

bors r-

@bors
Copy link
Contributor

bors bot commented Jan 15, 2022

Canceled.

DhairyaLGandhi
DhairyaLGandhi approved these changes Jan 15, 2022
View reviewed changes
Copy link
Member

@DhairyaLGandhi DhairyaLGandhi left a comment

Choose a reason for hiding this comment

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

bors r+

@ToucheSir
Copy link
Member Author

ToucheSir commented Jan 15, 2022

Testing whether this note from https://docs.nvidia.com/deeplearning/cudnn/release-notes/rel_8.html#rel-830 applies:

In some cases, cudnnConvolutionBackwardFilter() generates numerically imprecise results when used with algo set to CUDNN_CONVOLUTION_BWD_FILTER_ALGO_1. This issue is most frequently encountered with three dimensional spatial tensors. Users of the backend API may explicitly avoid backend engine 2032 or consider the numerical notes of engines and reject any marked as offering reduced precision reduction (CUDNN_NUMERICAL_NOTE_REDUCED_PRECISION_REDUCTION).

@bors
Copy link
Contributor

bors bot commented Jan 15, 2022

Build succeeded:

@bors bors bot merged commit 9c789dc into master Jan 15, 2022
This was referenced Jan 15, 2022
Closed
Closed
@ToucheSir ToucheSir deleted the bc/buildkite-latest-cuda branch January 15, 2022 23:53
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@DhairyaLGandhi DhairyaLGandhi DhairyaLGandhi approved these changes

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@ToucheSir @DhairyaLGandhi