Adapt DeepSeek persistent worker to offset cache copies - #118
Draft
high-cloud wants to merge 1 commit into
Draft
Conversation
|
Important Review skippedDraft detected. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Plus Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
copy_to.reset_persistent_windows=False, so serving does not perform a host-side comm-window reset between requests.Root cause
The latest simpler device-pointer provenance guard requires
copy_toto receive the base address of a live allocation. DeepSeek work-cache synchronization previously passedshard.data_ptr + offset, which is an interior pointer and is rejected before prefill starts.Dependencies
This PR depends on:
DistributedWorker.copy_to_offset(...);Orchestrator.copy_to_offset(...)and validates the live base allocation before applying the byte offset;reset_persistent_windows=Falsethe persistent-worker default.Validation
python -m pytest -q tests/test_deepseek_v4.py— 32 passed.ruff check --config ruff.toml pypto_serving/model/deepseek/npu_runner.py— passed.python tests/lint/check_headers.py— passed.python tests/lint/check_english_only.py— 53 passed.python -m compileall -q pypto_serving/model/deepseek/npu_runner.py— passed.task_20260727_032719_85546414704— 4 passed, serving shut down normally, and profiling trace was written toartifacts/forward-end-clear-mtp-v2/trace.json.