[TRTLLM-12112][feat] Support v2 KV cache stats#13953
Merged
lfr-0531 merged 10 commits intoMay 20, 2026
Merged
Conversation
yizhang-nv
force-pushed
the
feat/kv-cache-v2-stats-port
branch
7 times, most recently
from
c022723 to
c02889d
Compare
yizhang-nv
requested review from
dongxuy04,
hchings and
liji-nv
and removed request for
a team
yizhang-nv
force-pushed
the
feat/kv-cache-v2-stats-port
branch
from
c02889d to
4e40e05
Compare
yizhang-nv
force-pushed
the
feat/kv-cache-v2-stats-port
branch
8 times, most recently
from
c705ef1 to
5273874
Compare
lfr-0531
force-pushed
the
feat/deepseek_v4
branch
from
0a93d10 to
118e7a5
Compare
Collaborator
|
/bot run --disable-fail-fast |
Collaborator
|
PR_Github #48918 [ run ] triggered by Bot. Commit: |
lowsfer
reviewed
May 18, 2026
Collaborator
|
PR_Github #48918 [ run ] completed with state
|
yizhang-nv
changed the title
Member
Author
|
/bot run |
Collaborator
|
PR_Github #49133 [ run ] triggered by Bot. Commit: |
yizhang-nv
changed the title
yizhang-nv
changed the title
Member
Author
|
/bot run |
Collaborator
|
PR_Github #49147 [ run ] triggered by Bot. Commit: |
Collaborator
|
PR_Github #49133 [ run ] completed with state |
Collaborator
|
PR_Github #49147 [ run ] completed with state
|
Collaborator
|
@lowsfer could you review this? thanks |
Signed-off-by: Yi Zhang <187001205+yizhang-nv@users.noreply.github.com>
Signed-off-by: Yi Zhang <187001205+yizhang-nv@users.noreply.github.com>
Signed-off-by: Yi Zhang <187001205+yizhang-nv@users.noreply.github.com>
Signed-off-by: Yi Zhang <187001205+yizhang-nv@users.noreply.github.com>
Signed-off-by: Yi Zhang <187001205+yizhang-nv@users.noreply.github.com>
Signed-off-by: Yi Zhang <187001205+yizhang-nv@users.noreply.github.com>
Signed-off-by: Yi Zhang <187001205+yizhang-nv@users.noreply.github.com>
Signed-off-by: Yi Zhang <187001205+yizhang-nv@users.noreply.github.com>
Signed-off-by: Yi Zhang <187001205+yizhang-nv@users.noreply.github.com>
Signed-off-by: Yi Zhang <187001205+yizhang-nv@users.noreply.github.com>
yizhang-nv
force-pushed
the
feat/kv-cache-v2-stats-port
branch
from
bd74724 to
9ec657f
Compare
Member
Author
|
/bot run --disable-fail-fast |
Collaborator
|
PR_Github #49308 [ run ] triggered by Bot. Commit: |
Collaborator
|
PR_Github #49308 [ run ] completed with state |
lfr-0531
pushed a commit
to lfr-0531/TensorRT-LLM
that referenced
this pull request
May 29, 2026
Filtered out disaggregation and AutoDeploy-only changes for user/fanrongl/dsv4_model. Signed-off-by: Fanrong Li <lfr-0531@users.noreply.github.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
6 participants
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
Port V2 KV cache stats support onto the DeepSeek V4 feature branch.
This PR adds V2 KV cache stats delta tracking and wires it into the existing PyExecutor stats path so
LLM.get_stats()and server polling can surfacekvCacheStats/kvCacheIterationStatsfor KV cache manager V2. It also adds V1/V2 alignment coverage for basic block counts, SWA windows, and stacked multi-window reuse scenarios.Test Coverage
python3 -m py_compile tests/integration/defs/kv_cache/test_kv_cache_iteration_stats_alignment.pygit diff --check -- tests/integration/defs/kv_cache/test_kv_cache_iteration_stats_alignment.py/home/yizhan/.local/bin/pre-commit run --files tests/integration/defs/kv_cache/test_kv_cache_iteration_stats_alignment.pyLLM_MODELS_ROOT=/scratch.trt_llm_data/llm-models pytest -q -s tests/integration/defs/kv_cache/test_kv_cache_iteration_stats_alignment.py3 passed, 3 warnings in 153.30s/home/scratch.yizhan_sw_1/logs/2026-05-08/6u1g-0015/LLM_MODELS_ROOT__scratch_trt_llm_data_llm-models___01-35-50.stdout.logPR 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)
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.