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

[Paddle-TRT] Run radix sort of proposals layers on the context stream #31631

Merged
merged 1 commit into from
Mar 19, 2021
Merged

[Paddle-TRT] Run radix sort of proposals layers on the context stream #31631

merged 1 commit into from
Mar 19, 2021

Conversation

zlsh80826
Copy link
Collaborator

@zlsh80826 zlsh80826 commented Mar 15, 2021 •
edited
Loading

PR types

Performance optimization

PR changes

OPs

Describe

  1. Run radix sort of proposals layers on the context stream. Without indicating the stream, functions will be placed on default stream, which introduces additional implicit synchronization overhead.
  2. The is ~1.5% e2e performance gain on trt-fp16/maskrcnn inference

@cryoco cryoco requested a review from jerrywgz March 18, 2021 06:21
jerrywgz
jerrywgz reviewed Mar 19, 2021
View reviewed changes
Copy link
Contributor

@jerrywgz jerrywgz left a comment •
edited
Loading

Choose a reason for hiding this comment

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

./paddle/fluid/operators/detection/generate_proposals_v2_op.cu also uses SortDescending, which may applied this optimization as well?

@zlsh80826
Copy link
Collaborator Author

./paddle/fluid/operators/detection/generate_proposals_v2_op.cu also uses SortDescending, which may applied this optimization as well?

Yes, the modification is on bbox_utils.cu, which is used by generate_proposals and generate_proposals_v2

jerrywgz
jerrywgz approved these changes Mar 19, 2021
View reviewed changes
Copy link
Contributor

@jerrywgz jerrywgz left a comment

Choose a reason for hiding this comment

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

LGTM

@cryoco cryoco merged commit 1c67cf0 into PaddlePaddle:develop Mar 19, 2021
@zlsh80826 zlsh80826 changed the title Run radix sort of proposals layers on the context stream [Paddle-TRT] Run radix sort of proposals layers on the context stream Jan 6, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@jerrywgz jerrywgz jerrywgz approved these changes

Assignees
No one assigned
Labels
NVIDIA
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@zlsh80826 @jerrywgz @cryoco