[WIP] refactor(scheduler): unify hybrid cache model boundaries.#243
Open
SimonCqk wants to merge 2 commits into
Open
[WIP] refactor(scheduler): unify hybrid cache model boundaries.#243SimonCqk wants to merge 2 commits into
SimonCqk wants to merge 2 commits into
Conversation
SimonCqk
changed the title
Unify the scheduler-facing hybrid cache lifecycle around HybridPrefixCache so scheduler/FSM code calls lifecycle methods instead of spelling Mamba and paged-cache choreography directly. Add the family registry/recovery/lifecycle request types, RequestCacheContext, and focused diagnostics/stat helpers needed for the unified cache model. Split family-specific implementation into registry, Mamba, and paged-cache modules, while keeping concrete allocator semantics and avoiding a broad virtual family framework. Consolidate tests around durable facade behavior: request-local accounting, Mamba slot release/protection, worker metadata propagation, paged-cache matching, admission, eviction, and registry behavior. Signed-off-by: SimonCqk <cqk0100@gmail.com>
Signed-off-by: SimonCqk <cqk0100@gmail.com>
SimonCqk
force-pushed
the
feat/hybrid-cache-abstraction
branch
from
8590a0d to
0bb1c85
Compare
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Motivation
The scheduler path previously mixed scheduling logic with Mamba and paged-cache-specific lifecycle details. This refactor makes the scheduler depend on
a smaller cache lifecycle surface while keeping concrete family behavior inside the hybrid cache layer.
This is intentionally a concrete module split, not a broad virtual
CacheFamilyframework or universal allocator.Summary
HybridPrefixCachethe concrete lifecycle boundary forMatchPrefix,Admit,StepCommit,FinishRequest, andStats.family_registry,mamba_family_ops,paged_cache_family_opsTest Plan
69 passed210 passed