-
Notifications
You must be signed in to change notification settings - Fork 920
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.
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
Use the aligned_resource_adaptor
to allocate bloom filter device buffers
#17758
base: branch-25.02
Are you sure you want to change the base?
Use the aligned_resource_adaptor
to allocate bloom filter device buffers
#17758
Conversation
Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually. Contributors can view more details about this message here. |
aligned_resource_adaptor
to allocate bloom filter device buffers
aligned_resource_adaptor
to allocate bloom filter device buffersaligned_resource_adaptor
to allocate bloom filter device buffers
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good to me. It surprises me that this segfaulted the compiler in certain cases. We will need to file an internal bug report for this.
This PR doesn't unfortunately fix the compiler segfault unfortunately, (NVIDIA/cuCollections#660) should do that instead. I will update this PR to use the new |
aligned_resource_adaptor
to allocate bloom filter device buffersaligned_resource_adaptor
to allocate bloom filter device buffers
Locally verified with test data that the bloom filter is working as expected |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I tested locally with the 12.0 pip unified container and the compilation finished without issues.
@mhaseeb123 great work! And thank you, @sleeepyjack for the prompt fix on the cuco end.
Description
Related to #17164
Related to NVIDIA/cuCollections#660
This PR creates and uses a
rmm::mr::aligned_resource_adapter
to allocate device buffers for bloom filter data in accordance with bloom filter alignment requirements. This PR also updates thequery_bloom_filter
function to use the newbloom_filter_ref
constructors introduced in NVIDIA/cuCollections#660.Checklist