[None][feat] Merge back DeepSeek V4 cache manager updates#15633
Draft
jiaganc wants to merge 27 commits into
Draft
[None][feat] Merge back DeepSeek V4 cache manager updates#15633jiaganc wants to merge 27 commits into
jiaganc wants to merge 27 commits into
Conversation
Source-Commit: a7728aa Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
…ndexMapper release (NVIDIA#14423) Source-Commit: 02ac906 Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Source-Commit: 7a4b113 Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
NVIDIA#14308) Source-Commit: 2de7acb Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
…ength (NVIDIA#14627) Source-Commit: eeb09d1 Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Source-Commit: 69e7acc Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
…euse' + evictable gauges (NVIDIA#14544) Source-Commit: 772f750 Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
…s for safe order (NVIDIA#14674) Source-Commit: 808b24f Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
…_disagg_gen_init (NVIDIA#14695) Source-Commit: be4aa75 Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
…uests (+ probe hang-detector fix) (NVIDIA#14912) Source-Commit: 49f1ce6ce5d6847af7217e3cf5ba3ff9f5e2fdb2 Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Source-Commit: 95cd755 Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Source-Commit: 337b4a5 Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Source-Commit: 4f9d819 Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Source-Commit: c185066 Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Source-Commit: 0e22041 Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
…NVIDIA#15134) Source-Commit: ec3da87a9b688016f8f76cc43be9cf30c92c868e Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
…or (NVIDIA#15211) Source-Commit: c4cd713 Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Source-Commit: d15f21c Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Reason: rebase adaptation for PR1 final validation build failure. Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
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.
@coderabbitai summary
Description
Merge back the PR1 DeepSeek V4 cache-manager work onto current
main.This includes DeepSeek V4 sparse cache-manager updates, KVCacheManagerV2
adaptations after its migration from
resource_manager.pytokv_cache_manager_v2.py, disaggregated KV-transfer/cache-reuse updates,iteration stats reporting, OpenAI serving metrics updates, and rebase
compatibility fixes needed on top of current
main.This PR is opened as a draft because selected changed tests were intentionally
skipped by request and broader CI/build validation can still run before review.
Test Coverage
Remote unit tests were run on
umb-b300-dp-186inside containertrtllm-main-dsv4-cache-manager-mergeback-round2-jenkins-jiagancat SHAe188f8076c.Passed:
Result:
283 passed, 3 skipped, 4 warnings in 129.98sResult:
149 passed, 4 warnings in 71.21sResult:
392 passed, 66 skipped, 4 warnings in 146.77sResult:
63 passed, 12 skipped, 5 warnings in 338.84sSkipped by request:
tests/unittest/_torch/modeling/test_modeling_deepseekv4.pytests/unittest/_torch/attention/sparse/deepseek_v4/test_deepseek_v4_indices_transform.pytests/unittest/_torch/attention/sparse/deepseek_v4/test_deepseek_v4_sparse_mla.pytests/unittest/_torch/speculative/test_eagle3.pyPR Checklist
Please review the following before submitting your PR:
PR description clearly explains what and why. If using CodeRabbit's summary, please make sure it makes sense.
PR Follows TRT-LLM CODING GUIDELINES to the best of your knowledge.
Test cases are provided for new code paths (see test instructions)
If PR introduces API changes, an appropriate PR label is added - either
api-compatibleorapi-breaking. Forapi-breaking, includeBREAKINGin the PR title.Any new dependencies have been scanned for license and vulnerabilities
CODEOWNERS updated if ownership changes
Documentation updated as needed
Update tava architecture diagram if there is a significant design change in PR.
The reviewers assigned automatically/manually are appropriate for the PR.
Please check this after reviewing the above items as appropriate for this PR.
GitHub Bot Help
To see a list of available CI bot commands, please comment
/bot help.