Add Bloom filter persisting L2 access#826
Merged
Merged
Conversation
Collaborator
Author
|
/ok to test 54e5616 |
sleeepyjack
force-pushed
the
bloom-filter-perf-study
branch
from
54e5616 to
b6ba6be
Compare
sleeepyjack
changed the title
Rename the Bloom filter policy/header, make the policy key-based with an explicit hash parameter, and use it directly as the default policy. Default the policy word/block shape, vectorization layouts, and tuning booleans.
Add optional persisting-L2 filter accesses using associate_access_property, with tests for global-memory equivalence and shared-memory safety. Add a documented CUDA runtime example showing L2 set-aside and reset usage.
sleeepyjack
commented
Jul 3, 2026
Comment on lines
+46
to
+53
| * NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition, 128 MiB L2, 200M input items: | ||
| * | ||
| * filter size | add, default | add, persisting L2 | contains, default | contains, persisting L2 | ||
| * ----------- | ------------ | ------------------ | ----------------- | ------------------------ | ||
| * 64 MiB | 4.880 ms | 3.405 ms | 2.444 ms | 2.427 ms | ||
| * 80 MiB | 8.276 ms | 3.380 ms | 4.706 ms | 2.443 ms | ||
| * 96 MiB | 11.004 ms | 3.476 ms | 4.874 ms | 2.575 ms | ||
| * 128 MiB | 22.602 ms | 9.097 ms | 5.545 ms | 5.371 ms |
Collaborator
Author
There was a problem hiding this comment.
Benchmark results
sleeepyjack
added
topic: performance
Collaborator
Author
|
/ok to test 90517ea |
PointKernel
approved these changes
Jul 6, 2026
PointKernel
left a comment
PointKernel
left a comment
Member
There was a problem hiding this comment.
Looks good. Just some small non-blocking questions
|
this is amazing! |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Depends on #827.
cuda::associate_access_propertyaddandcontainsadd/contains, then resets the set-aside