Lead the README with SparkCache behavior - #24
Closed
FujitsuPolycom wants to merge 3 commits into
Closed
Conversation
Describe persistent prefix reuse, rank-local restore, and verified recomputation before artifact-specific qualification details. Move package and OCI evidence beside deployment guidance, add implemented tail-publication and row-segment-sharing capabilities, distinguish their model applicability and qualification state, and list the e10536a and 0b67266 runtime contracts. SparkCache wire values, cache identity, and runtime behavior are unchanged. Validation: 738 SparkCache tests passed with 7 skipped; 108 deployment tests passed with 1 skipped; Ruff and Markdown diff checks passed.
Replace ambiguous native-placement wording with SparkCache C++/CUDA page restore and name Python/Torch page restore as the comparison path. Clarify that SparkCache performs verified transfer into vLLM-owned cache pages. Documentation only; cache identity and runtime behavior are unchanged. Validation: Markdown diff checks passed.
Replace the lifecycle warning label with an exact qualification boundary and describe subsequent-request reuse without temporal shorthand. Documentation only; cache identity and runtime behavior are unchanged. Validation: Markdown diff checks passed.
Owner
Author
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Resulting documentation
The README opens with SparkCache's purpose: persistent rank-local prefix reuse, local restore, and verified recomputation when stored state cannot be proven compatible. Package, OCI, and source-deployment qualification details appear beside deployment guidance.
The capability table includes the implemented tail-only row and opaque-page publication paths plus different-root row-segment sharing. It distinguishes exact-prefix GPU reuse from different-root sharing, limits SparkCache C++/CUDA restore qualification to the recorded flat-snapshot runtime, and identifies capabilities without live qualification or applicability to opaque GLM pages.
The vLLM source-contract table records the e10536a adaptive-MTP contract and the 0b67266 adaptive-MTP/live-tensor-KDA contract.
Compatibility
Documentation only. SparkCache wire values, cache identity, and runtime behavior are unchanged.
Validation