Skip to content

Conversation

@DwyaneShi
Copy link
Collaborator

Pull Request Description

This PR includes core modules of L2Cache.

  • connectors to support different L2Cache backends
  • key builders
  • marshallers
  • key builder benchmark

Related Issues

Resolves: #[Insert issue number(s)]

Important: Before submitting, please complete the description above and review the checklist below.


Contribution Guidelines (Expand for Details)

We appreciate your contribution to aibrix! To ensure a smooth review process and maintain high code quality, please adhere to the following guidelines:

Pull Request Title Format

Your PR title should start with one of these prefixes to indicate the nature of the change:

  • [Bug]: Corrections to existing functionality
  • [CI]: Changes to build process or CI pipeline
  • [Docs]: Updates or additions to documentation
  • [API]: Modifications to aibrix's API or interface
  • [CLI]: Changes or additions to the Command Line Interface
  • [Misc]: For changes not covered above (use sparingly)

Note: For changes spanning multiple categories, use multiple prefixes in order of importance.

Submission Checklist

  • PR title includes appropriate prefix(es)
  • Changes are clearly explained in the PR description
  • New and existing tests pass successfully
  • Code adheres to project style and best practices
  • Documentation updated to reflect changes (if applicable)
  • Thorough testing completed, no regressions introduced

By submitting this PR, you confirm that you've read these guidelines and your changes align with the project's contribution standards.

@DwyaneShi DwyaneShi force-pushed the haiyang/kvcache-l2cache-part1 branch from 79f0699 to a8833d0 Compare May 7, 2025 18:05
- connectors
- key builders
- marshallers
- key builder benchmark

Signed-off-by: Haiyang Shi <[email protected]>
@DwyaneShi DwyaneShi force-pushed the haiyang/kvcache-l2cache-part1 branch from a8833d0 to 17a2b7c Compare May 7, 2025 18:13
@DwyaneShi DwyaneShi requested review from Jeffwan and happyandslow May 7, 2025 18:15
redis = "^6.0.0"
fakeredis = "^2.28.1"
codespell = "2.4.1"
infinistore = "^0.2.35"
Copy link
Collaborator

Choose a reason for hiding this comment

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

does hpkv provide whl in pypi ?

Copy link
Collaborator Author

Choose a reason for hiding this comment

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

HPKV is provided in our internal pypi, not available from the public one

# index to support multi-GPU per RNIC. For example, if we have 8
# GPUs and 2 RNICs, then GPU 0 to 3 are mapped to RNIC 0, and GPU
# 4 to 7 are mapped to RNIC 1.
factor = num_visible_gpus // len(dev_list)
Copy link
Collaborator

Choose a reason for hiding this comment

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

em. we may not explicitly pass device list to engine for TP <= 8 case. we can do some improvements later to intelligently search dev_list if that's not given from user.

Copy link
Collaborator Author

Choose a reason for hiding this comment

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

sure

gpu_idx = torch.cuda.current_device()
rnic_idx = gpu_idx // factor
dev_name = dev_list[rnic_idx]

Copy link
Collaborator

Choose a reason for hiding this comment

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

TODO: add gid search logic and add --hint-gid-index to infinistore.ClientConfig to support VNI mode, this can be done later

@DwyaneShi DwyaneShi merged commit 961eeed into vllm-project:main May 7, 2025
13 checks passed
Yaegaki1Erika pushed a commit to Yaegaki1Erika/aibrix that referenced this pull request Jul 23, 2025
- connectors
- key builders
- marshallers
- key builder benchmark

Signed-off-by: Haiyang Shi <[email protected]>
Co-authored-by: Haiyang Shi <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants