[None][DO NOT REVIEW] Trigger CI Only#11463
Conversation
yizhang-nv
force-pushed
the
enable-v2-by-default
branch
from
1de7c1a to
9c0eb21
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #35703 [ run ] triggered by Bot. Commit: |
|
/bot run --disable-fail-fast |
|
PR_Github #35725 [ run ] triggered by Bot. Commit: |
|
PR_Github #35725 [ run ] completed with state
|
yizhang-nv
force-pushed
the
enable-v2-by-default
branch
from
c63b434 to
39d0d39
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #36446 [ run ] triggered by Bot. Commit: |
|
PR_Github #36446 [ run ] completed with state
|
|
/bot run --disable-fail-fast |
|
PR_Github #36493 [ run ] triggered by Bot. Commit: |
yizhang-nv
force-pushed
the
enable-v2-by-default
branch
from
dbffd28 to
fddeb28
Compare
|
/bot run --disable-fail-fast |
1 similar comment
|
/bot run --disable-fail-fast |
yizhang-nv
force-pushed
the
enable-v2-by-default
branch
from
fddeb28 to
2b7c494
Compare
|
/bot kill |
|
PR_Github #36497 [ run ] triggered by Bot. Commit: |
|
PR_Github #36499 [ run ] triggered by Bot. Commit: |
|
PR_Github #36500 [ kill ] triggered by Bot. Commit: |
|
PR_Github #36499 [ run ] completed with state |
|
PR_Github #36500 [ kill ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #36501 [ run ] triggered by Bot. Commit: |
|
PR_Github #36501 [ run ] completed with state
|
yizhang-nv
force-pushed
the
enable-v2-by-default
branch
from
2b7c494 to
44b6a1a
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #36591 [ run ] triggered by Bot. Commit: |
|
/bot run --help |
|
/bot run --disable-fail-fast --post-merge |
|
PR_Github #38293 [ run ] completed with state
|
yizhang-nv
force-pushed
the
enable-v2-by-default
branch
from
b03fb89 to
cd2f405
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #38341 [ run ] triggered by Bot. Commit: |
|
PR_Github #38341 [ run ] completed with state
|
Switch kv_cache_memory_pool_version default from 1 to 2 and waive tests that fail under v2 due to OOM or unsupported features. Signed-off-by: Yi Zhang <yizhan@nvidia.com> Signed-off-by: Yi Zhang <187001205+yizhang-nv@users.noreply.github.com>
Skip tests that fail due to KV cache memory pool v2 init OOM on draft model KV cache allocation (Eagle3/PARD/MTP speculative decoding) and other CI-only failures unrelated to v2. - 10 accuracy tests: CuOOMError in kv_cache_manager_v2 during draft model KV cache init on DGX_H100, DGX_B200, B300 - 4 perf sanity tests: DeepSeek R1 FP4 v2 + MTP3 on GB200 - test_visual_gen_quickstart: CUDA OOM on A10/RTX5080 (possibly related) - test_openai_lora, test_trtllm_serve_lora_example: local pass, CI fail - test_openai_chat_multimodal_example: possibly precision issue Signed-off-by: Yi Zhang <yizhang@nvidia.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
enable-v2-by-default
branch
from
cd2f405 to
f498a1c
Compare
|
/bot run |
|
PR_Github #38553 [ run ] triggered by Bot. Commit: |
|
/bot run --disable-fail-fast |
1 similar comment
|
/bot run --disable-fail-fast |
|
PR_Github #38555 [ run ] triggered by Bot. Commit: |
|
PR_Github #38556 [ run ] triggered by Bot. Commit: |
|
PR_Github #38555 [ run ] completed with state |
PR NVIDIA#11907 changed KVCacheManagerV2 to use max_gpu_total_bytes instead of max_tokens for capacity control. However, target and draft KV cache managers share the same kv_cache_config. With max_tokens=None, both managers take the full max_gpu_total_bytes (~91 GiB) as their quota, causing OOM at init when the combined allocation exceeds GPU memory. Root cause: - configure_kv_cache_capacity computes a total budget for target + draft - PR NVIDIA#11907 writes this total to max_gpu_total_bytes in the shared config - Both target and draft V2 managers each try to allocate the full budget - V2 pre-allocates via CUDA virtual memory, so quota > available = OOM Fix: 1. Split max_gpu_total_bytes proportionally between target and draft based on their per-token KV cache sizes before creating the final managers. Draft gets a cloned KvCacheConfig with its share of the budget. 2. Fix get_cache_size_per_token to accept an optional num_layers parameter. For draft models (EAGLE3/MTP), the HF config num_hidden_layers (e.g. 1) differs from the actual runtime layer count (e.g. 4 for EAGLE3). This caused _get_kv_size_per_token to underestimate draft cost by 4x, making both V1 and V2 allocate more than the total budget in aggregate. 3. Make _get_kv_size_per_token PP-aware for draft layers: EAGLE3/MTP draft layers are only on the last PP rank, so only that rank includes draft cost in its per-token estimate. Tested on B200 (183 GiB): - TestQwen3_8B::test_eagle3: PASSED (was OOM) - TestLlama3_1_8BInstruct::test_guided_decoding_with_eagle3: PASSED - TestDeepSeekV3Lite::test_bfloat16_python_scheduler[mtp=2]: PASSED - TestDeepSeekV3Lite::test_nvfp4[mtp=2]: PASSED - TestDeepSeekV3Lite::test_guided_decoding[mtp=2]: PASSED - test_llmapi_speculative_decoding_mtp: PASSED - test_llmapi_speculative_decoding_eagle3: PASSED - test_visual_gen_quickstart: PASSED (B200), FAILED (A10 — V1 also OOM) Signed-off-by: Yi Zhang <187001205+yizhang-nv@users.noreply.github.com>
|
PR_Github #38556 [ run ] completed with state
|
|
/bot run --disable-fail-fast |
|
PR_Github #38602 [ run ] triggered by Bot. Commit: |
|
PR_Github #38602 [ run ] completed with state
|
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
|
/bot run ci |
|
/bot run ci disable fail fast |
|
/bot run --disable-fail-fast |
|
PR_Github #38658 Bot args parsing error: usage: /bot [-h] |
|
PR_Github #38659 [ run ] triggered by Bot. Commit: |
|
/bot --help |
GitHub Bot Help
Provide a user friendly way for developers to interact with a Jenkins server. Run See details below for each supported subcommand. Details
Launch build/test pipelines. All previously running jobs will be killed.
kill
Kill all running builds associated with pull request. skip
Skip testing for latest commit on pull request. reuse-pipeline
Reuse a previous pipeline to validate current commit. This action will also kill all currently running builds associated with the pull request. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break. |
The OOM issue during draft model KV cache allocation was fixed in 50a2f87, so these test cases no longer need to be skipped. Signed-off-by: Yi Zhang <187001205+yizhang-nv@users.noreply.github.com>
2 participants
@coderabbitai summary
Description
Test Coverage
PR 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
/bot [-h] ['run', 'kill', 'skip', 'reuse-pipeline'] ...Provide a user friendly way for developers to interact with a Jenkins server.
Run
/bot [-h|--help]to print this help message.See details below for each supported subcommand.
Details
run [--reuse-test (optional)pipeline-id --disable-fail-fast --skip-test --stage-list "A10-PyTorch-1, xxx" --gpu-type "A30, H100_PCIe" --test-backend "pytorch, cpp" --add-multi-gpu-test --only-multi-gpu-test --disable-multi-gpu-test --post-merge --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx" --detailed-log --debug(experimental)]Launch build/test pipelines. All previously running jobs will be killed.
--reuse-test (optional)pipeline-id(OPTIONAL) : Allow the new pipeline to reuse build artifacts and skip successful test stages from a specified pipeline or the last pipeline if no pipeline-id is indicated. If the Git commit ID has changed, this option will be always ignored. The DEFAULT behavior of the bot is to reuse build artifacts and successful test results from the last pipeline.--disable-reuse-test(OPTIONAL) : Explicitly prevent the pipeline from reusing build artifacts and skipping successful test stages from a previous pipeline. Ensure that all builds and tests are run regardless of previous successes.--disable-fail-fast(OPTIONAL) : Disable fail fast on build/tests/infra failures.--skip-test(OPTIONAL) : Skip all test stages, but still run build stages, package stages and sanity check stages. Note: Does NOT update GitHub check status.--stage-list "A10-PyTorch-1, xxx"(OPTIONAL) : Only run the specified test stages. Examples: "A10-PyTorch-1, xxx". Note: Does NOT update GitHub check status.--gpu-type "A30, H100_PCIe"(OPTIONAL) : Only run the test stages on the specified GPU types. Examples: "A30, H100_PCIe". Note: Does NOT update GitHub check status.--test-backend "pytorch, cpp"(OPTIONAL) : Skip test stages which don't match the specified backends. Only support [pytorch, cpp, tensorrt, triton]. Examples: "pytorch, cpp" (does not run test stages with tensorrt or triton backend). Note: Does NOT update GitHub pipeline status.--only-multi-gpu-test(OPTIONAL) : Only run the multi-GPU tests. Note: Does NOT update GitHub check status.--disable-multi-gpu-test(OPTIONAL) : Disable the multi-GPU tests. Note: Does NOT update GitHub check status.--add-multi-gpu-test(OPTIONAL) : Force run the multi-GPU tests in addition to running L0 pre-merge pipeline.--post-merge(OPTIONAL) : Run the L0 post-merge pipeline instead of the ordinary L0 pre-merge pipeline.--extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx"(OPTIONAL) : Run the ordinary L0 pre-merge pipeline and specified test stages. Examples: --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx".--detailed-log(OPTIONAL) : Enable flushing out all logs to the Jenkins console. This will significantly increase the log volume and may slow down the job.--debug(OPTIONAL) : Experimental feature. Enable access to the CI container for debugging purpose. Note: Specify exactly one stage in thestage-listparameter to access the appropriate container environment. Note: Does NOT update GitHub check status.For guidance on mapping tests to stage names, see
docs/source/reference/ci-overview.mdand the
scripts/test_to_stage_mapping.pyhelper.kill
killKill all running builds associated with pull request.
skip
skip --comment COMMENTSkip testing for latest commit on pull request.
--comment "Reason for skipping build/test"is required. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.reuse-pipeline
reuse-pipelineReuse a previous pipeline to validate current commit. This action will also kill all currently running builds associated with the pull request. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.