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

Re-evaluate the prefetching fallback kernel and compare it with DeviceFor as fallback #2363

Closed
4 tasks done
bernhardmgruber opened this issue Sep 4, 2024 · 0 comments · Fixed by #2396
Closed
4 tasks done

Re-evaluate the prefetching fallback kernel and compare it with DeviceFor as fallback #2363

bernhardmgruber opened this issue Sep 4, 2024 · 0 comments · Fixed by #2396
Assignees
@bernhardmgruber
Labels
cub For all items related to CUB

Comments

@bernhardmgruber
Copy link
Contributor

bernhardmgruber commented Sep 4, 2024 •
edited
Loading

The inital PR for cub::DeviceTransform #2086 had a dedicated fallback kernel using prefetching instead of falling back to cub::DeviceFor. This kernel was removed during the review, to make way for the performance validated Hopper kernel. We should

  • properly validate the prefetching kernel's performance
  • bring it back if it offers enough benefit
  • improve it to beat cub::DeviceFor
  • replace cub::DeviceFor as fallback kernel for cub::DeviceTransform.
@github-project-automation github-project-automation bot moved this to Todo in CCCL Sep 4, 2024
@bernhardmgruber bernhardmgruber self-assigned this Sep 4, 2024
@bernhardmgruber bernhardmgruber added the cub For all items related to CUB label Sep 4, 2024
bernhardmgruber added a commit to bernhardmgruber/cccl that referenced this issue Sep 10, 2024
@bernhardmgruber
Add prefetch transform kernel
0cfe4ba 
Fixes: NVIDIA#2363
@bernhardmgruber bernhardmgruber mentioned this issue Sep 10, 2024
Merged
3 tasks
@cccl-authenticator-app cccl-authenticator-app bot moved this from Todo to In Progress in CCCL Sep 10, 2024
bernhardmgruber added a commit to bernhardmgruber/cccl that referenced this issue Sep 10, 2024
@bernhardmgruber
Add prefetch transform kernel
74fb7d3 
Fixes: NVIDIA#2363
@cccl-authenticator-app cccl-authenticator-app bot moved this from In Progress to In Review in CCCL Sep 10, 2024
@bernhardmgruber bernhardmgruber mentioned this issue Sep 6, 2024
Open
19 tasks
copy-pr-bot bot pushed a commit that referenced this issue Oct 28, 2024
@bernhardmgruber
Add prefetch transform kernel
1b337c0 
Fixes: #2363
bernhardmgruber added a commit to bernhardmgruber/cccl that referenced this issue Oct 29, 2024
@bernhardmgruber
Add prefetch transform kernel
5da9daf 
Fixes: NVIDIA#2363
bernhardmgruber added a commit to bernhardmgruber/cccl that referenced this issue Oct 29, 2024
@bernhardmgruber
Add prefetch transform kernel
ff90e90 
Fixes: NVIDIA#2363
bernhardmgruber added a commit to bernhardmgruber/cccl that referenced this issue Oct 30, 2024
@bernhardmgruber
Add prefetch transform kernel
33ced85 
Fixes: NVIDIA#2363
bernhardmgruber added a commit to bernhardmgruber/cccl that referenced this issue Oct 30, 2024
@bernhardmgruber
Add prefetch transform kernel
65605a2 
Fixes: NVIDIA#2363
bernhardmgruber added a commit that referenced this issue Oct 30, 2024
@bernhardmgruber
Add prefetching kernel as new fallback for cub::DeviceTransform (#2396
2f05ef3 
)

Fixes: #2363
@github-project-automation github-project-automation bot moved this from In Review to Done in CCCL Oct 30, 2024
fbusato pushed a commit to fbusato/cccl that referenced this issue Nov 5, 2024
@bernhardmgruber @fbusato
Add prefetching kernel as new fallback for cub::DeviceTransform (NV…
84e09c1 
…IDIA#2396)

Fixes: NVIDIA#2363
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@bernhardmgruber bernhardmgruber

Labels
cub For all items related to CUB
Projects
CCCL
Archived in project
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

Add prefetching kernel as new fallback for cub::DeviceTransform bernhardmgruber/cccl
1 participant
@bernhardmgruber