[TRTLLM-5972][chore] Load balance decode token KV cache with helix parallelism #9757
Conversation
brb-nv
force-pushed
the
user/brb/load-balance-helix-decode-tokens
branch
2 times, most recently
from
4839014 to
6b60df1
Compare
📝 WalkthroughWalkthroughThe changes enhance CP Helix generation by adding per-rank sequence length tracking, refining past token count computation based on rank status, and implementing round-robin decode block distribution across CP ranks to control which ranks process tokens during generation. Changes
Sequence DiagramsequenceDiagram
participant ReqQueue as Request Queue
participant ResMgr as Resource Manager
participant ModelEngine as Model Engine
ReqQueue->>ReqQueue: Set req.seqlen_this_rank_cp = len(input_ids_this_rank)
ResMgr->>ResMgr: Compute decode_block_id from generation iteration
rect rgb(200, 220, 255)
Note over ResMgr: Round-robin activation logic
alt decode_block_id % cp_size == cp_rank
ResMgr->>ResMgr: req.py_helix_is_inactive_rank = False
ResMgr->>ResMgr: Increment seqlen_this_rank_cp
else
ResMgr->>ResMgr: req.py_helix_is_inactive_rank = True
ResMgr->>ResMgr: Skip KV cache allocation
end
end
ModelEngine->>ModelEngine: Check py_helix_is_inactive_rank
rect rgb(220, 240, 220)
Note over ModelEngine: Past token computation
alt Rank is inactive
ModelEngine->>ModelEngine: past_seen_token_num = seqlen_this_rank_cp
else
ModelEngine->>ModelEngine: past_seen_token_num = seqlen_this_rank_cp - 1
end
end
ModelEngine->>ModelEngine: Compute position_id and token lengths
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes
Pre-merge checks and finishing touches❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✨ Finishing touches
🧪 Generate unit tests (beta)
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 |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Nitpick comments (1)
tensorrt_llm/_torch/pyexecutor/executor_request_queue.py (1)
687-701: Ensureseqlen_this_rank_cpis correctly set for all Helix LlmRequest instances (including beams).Initializing
req.seqlen_this_rank_cp = len(input_ids_this_rank)at prefill is correct and matches its later use as “tokens with KV on this CP rank.” Please double‑check that:
- Any
req.child_requestscreated byexecutor_request_to_llm_requestfor beam search also get a validseqlen_this_rank_cp(either via the helper or by explicitly copying the parent’s value here), and- No Helix generation request reaches
model_engine._prepare_tp_inputs/KVCacheManager.prepare_resourceswithout this field set.Otherwise, beam children or special paths could see an AttributeError or inconsistent per‑rank KV lengths at decode.
If needed, one simple defensive option is:
req.total_input_len_cp = input_len - req.seqlen_this_rank_cp = len(input_ids_this_rank) + req.seqlen_this_rank_cp = len(input_ids_this_rank) + if req.child_requests: + for child in req.child_requests: + setattr(child, "seqlen_this_rank_cp", + getattr(child, "seqlen_this_rank_cp", + req.seqlen_this_rank_cp))
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (3)
tensorrt_llm/_torch/pyexecutor/executor_request_queue.py(1 hunks)tensorrt_llm/_torch/pyexecutor/model_engine.py(1 hunks)tensorrt_llm/_torch/pyexecutor/resource_manager.py(1 hunks)
🧰 Additional context used
📓 Path-based instructions (2)
**/*.py
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
**/*.py: The code developed for TensorRT-LLM should conform to Python 3.8+
Indent Python code with 4 spaces; do not use tabs
Always maintain the namespace when importing in Python, even if only one class or function from a module is used (e.g., usefrom package.subpackage import fooand thenfoo.SomeClass()instead offrom package.subpackage.foo import SomeClass)
Python filenames should use snake_case (e.g.,some_file.py)
Python class names should use PascalCase (e.g.,class SomeClass)
Python function and method names should use snake_case (e.g.,def my_awesome_function():)
Python local variable names should use snake_case, with prefixkfor variable names that start with a number (e.g.,k_99th_percentile = ...)
Python global variables should use upper snake_case with prefixG(e.g.,G_MY_GLOBAL = ...)
Python constants should use upper snake_case (e.g.,MY_CONSTANT = ...)
Avoid shadowing variables declared in an outer scope in Python
Initialize all externally visible members of a Python class in the constructor
For Python interfaces that may be used outside a file, prefer docstrings over comments
Python comments should be reserved for code within a function, or interfaces that are local to a file
Use Google style docstrings for Python classes and functions, which can be parsed by Sphinx
Python attributes and variables can be documented inline with type and description (e.g.,self.x = 5followed by"""<type>: Description of 'x'""")
Avoid using reflection in Python when functionality can be easily achieved without reflection
When using try-except blocks in Python, limit the except clause to the smallest set of specific errors possible instead of catching all exceptions
When using try-except blocks in Python to handle multiple possible variable types (duck-typing), keep the body of the try as small as possible and use the else block to implement the logic
Files:
tensorrt_llm/_torch/pyexecutor/model_engine.pytensorrt_llm/_torch/pyexecutor/executor_request_queue.pytensorrt_llm/_torch/pyexecutor/resource_manager.py
**/*.{cpp,h,cu,py}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
All TensorRT-LLM Open Source Software code files should contain an NVIDIA copyright header that includes the current year at the top
Files:
tensorrt_llm/_torch/pyexecutor/model_engine.pytensorrt_llm/_torch/pyexecutor/executor_request_queue.pytensorrt_llm/_torch/pyexecutor/resource_manager.py
🧠 Learnings (14)
📓 Common learnings
Learnt from: thorjohnsen
Repo: NVIDIA/TensorRT-LLM PR: 6910
File: cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp:0-0
Timestamp: 2025-08-14T21:04:50.248Z
Learning: In KV cache onboarding logic during prefill in cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp, when calculating which blocks fall within the attention window, use getTokensPerBlock() to advance token indices rather than block->getUniqueTokens().size(), because the calculation needs to consider the post-prefill state where blocks will be filled to capacity, not their current token count.
Learnt from: nv-lschneider
Repo: NVIDIA/TensorRT-LLM PR: 7910
File: cpp/tensorrt_llm/kernels/nccl_device/config.cu:42-49
Timestamp: 2025-09-23T14:58:05.372Z
Learning: In TensorRT-LLM NCCL device kernels (cpp/tensorrt_llm/kernels/nccl_device/), the token partitioning intentionally uses ceil-like distribution (same token_per_rank for all ranks) to ensure all ranks launch the same number of blocks. This is required for optimal NCCL device API barrier performance, even though it may launch extra blocks for non-existent tokens on later ranks. Runtime bounds checking in the kernel (blockID validation) handles the overshoot cases.
Learnt from: eopXD
Repo: NVIDIA/TensorRT-LLM PR: 6767
File: cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp:0-0
Timestamp: 2025-08-15T06:46:54.897Z
Learning: In cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp addToken function, newly allocated blocks are unshared by design. The beam search path in addToken (when sequence.getNumTokens() > windowSize) is currently broken/non-functional with SWA, so the block allocation doesn't follow a shared-then-unshared pattern.
📚 Learning: 2025-08-19T12:45:11.997Z
Learnt from: amitz-nv
Repo: NVIDIA/TensorRT-LLM PR: 7033
File: tensorrt_llm/_torch/pyexecutor/model_engine.py:0-0
Timestamp: 2025-08-19T12:45:11.997Z
Learning: In tensorrt_llm/_torch/pyexecutor/model_engine.py, DoRA (Delta Orthogonal Rank Adaptation) functionality was removed from the PyTorch flow to eliminate issues with inverted DoRA detection logic. The original is_dora condition was checking if scaling_vec_pointer == 0, which was potentially incorrect.
Applied to files:
tensorrt_llm/_torch/pyexecutor/model_engine.pytensorrt_llm/_torch/pyexecutor/resource_manager.py
📚 Learning: 2025-09-23T14:58:05.372Z
Learnt from: nv-lschneider
Repo: NVIDIA/TensorRT-LLM PR: 7910
File: cpp/tensorrt_llm/kernels/nccl_device/config.cu:42-49
Timestamp: 2025-09-23T14:58:05.372Z
Learning: In TensorRT-LLM NCCL device kernels (cpp/tensorrt_llm/kernels/nccl_device/), the token partitioning intentionally uses ceil-like distribution (same token_per_rank for all ranks) to ensure all ranks launch the same number of blocks. This is required for optimal NCCL device API barrier performance, even though it may launch extra blocks for non-existent tokens on later ranks. Runtime bounds checking in the kernel (blockID validation) handles the overshoot cases.
Applied to files:
tensorrt_llm/_torch/pyexecutor/model_engine.pytensorrt_llm/_torch/pyexecutor/resource_manager.py
📚 Learning: 2025-08-26T06:07:02.166Z
Learnt from: shaharmor98
Repo: NVIDIA/TensorRT-LLM PR: 7231
File: tensorrt_llm/_torch/pyexecutor/_util.py:504-509
Timestamp: 2025-08-26T06:07:02.166Z
Learning: In tensorrt_llm/_torch/pyexecutor/_util.py, when calling model_engine.set_lora_model_config(), pass model_binding_config.mlp_hidden_size directly without multiplying by mapping.tp_size, as the mlp_hidden_size from get_bindings_model_config() is already the per-TP rank value needed for LoRA weight packaging.
Applied to files:
tensorrt_llm/_torch/pyexecutor/model_engine.pytensorrt_llm/_torch/pyexecutor/executor_request_queue.py
📚 Learning: 2025-08-14T21:04:50.248Z
Learnt from: thorjohnsen
Repo: NVIDIA/TensorRT-LLM PR: 6910
File: cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp:0-0
Timestamp: 2025-08-14T21:04:50.248Z
Learning: In KV cache onboarding logic during prefill in cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp, when calculating which blocks fall within the attention window, use getTokensPerBlock() to advance token indices rather than block->getUniqueTokens().size(), because the calculation needs to consider the post-prefill state where blocks will be filled to capacity, not their current token count.
Applied to files:
tensorrt_llm/_torch/pyexecutor/model_engine.pytensorrt_llm/_torch/pyexecutor/resource_manager.py
📚 Learning: 2025-08-09T20:57:04.084Z
Learnt from: sklevtsov-nvidia
Repo: NVIDIA/TensorRT-LLM PR: 3294
File: cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_tma_warp_specialized_input.cu:118-127
Timestamp: 2025-08-09T20:57:04.084Z
Learning: In the CUTLASS MoE finalize fusion implementation (cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_gemm_tma_warp_specialized_input.cu), when setting `fused_finalize_epilogue.stride_final_output` with shape `(hidden_size, num_output_tokens, 1)`, the `num_rows_in_final_output` should be set to `num_output_tokens` (not `hidden_size`) because of a swap+transpose operation that maps rows of the output tensor to `hidden_size` and columns to `num_output_tokens`.
Applied to files:
tensorrt_llm/_torch/pyexecutor/model_engine.py
📚 Learning: 2025-08-19T12:45:35.429Z
Learnt from: amitz-nv
Repo: NVIDIA/TensorRT-LLM PR: 7033
File: tensorrt_llm/_torch/pyexecutor/model_engine.py:2086-2092
Timestamp: 2025-08-19T12:45:35.429Z
Learning: DoRA (Delta Orthogonal Rank Adaptation) functionality has been removed from the PyTorch flow in tensorrt_llm/_torch/pyexecutor/model_engine.py. The is_dora field is computed but not used downstream in the PyTorch flow, so converting it to a tensor would be wasteful overhead.
Applied to files:
tensorrt_llm/_torch/pyexecutor/model_engine.py
📚 Learning: 2025-08-14T06:36:40.701Z
Learnt from: timlee0212
Repo: NVIDIA/TensorRT-LLM PR: 6886
File: tensorrt_llm/_torch/models/modeling_deepseekv3.py:0-0
Timestamp: 2025-08-14T06:36:40.701Z
Learning: In DeepSeek V3 model (tensorrt_llm/_torch/models/modeling_deepseekv3.py), the disagreement between AllReduce.__init__ guard and _compute_mlp_tp_size logic for MNNVL usage is expected by design. The AllReduce component and MLP TP-size computation intentionally use different criteria for MNNVL availability decisions.
Applied to files:
tensorrt_llm/_torch/pyexecutor/model_engine.py
📚 Learning: 2025-09-29T15:14:28.503Z
Learnt from: amitz-nv
Repo: NVIDIA/TensorRT-LLM PR: 8063
File: tensorrt_llm/lora_manager.py:1080-1112
Timestamp: 2025-09-29T15:14:28.503Z
Learning: In tensorrt_llm/lora_manager.py, when calculating part_sizes for attn_qkv fused LoRA modules, the sizes are correctly multiplied by tp_size because model_config.num_heads and model_config.num_kv_heads are already divided by tp_size (per-TP-rank values), so multiplication is needed to get the original full concatenated dimension size. The interleave_fused_lora_weights_for_tp function provides proper validation.
Applied to files:
tensorrt_llm/_torch/pyexecutor/model_engine.py
📚 Learning: 2025-08-18T08:42:02.640Z
Learnt from: samuellees
Repo: NVIDIA/TensorRT-LLM PR: 6974
File: tensorrt_llm/serve/scripts/benchmark_dataset.py:558-566
Timestamp: 2025-08-18T08:42:02.640Z
Learning: In TensorRT-LLM's RandomDataset (tensorrt_llm/serve/scripts/benchmark_dataset.py), when using --random-token-ids option, sequence length accuracy is prioritized over semantic correctness for benchmarking purposes. The encode/decode operations should use skip_special_tokens=True and add_special_tokens=False to ensure exact target token lengths.
Applied to files:
tensorrt_llm/_torch/pyexecutor/model_engine.py
📚 Learning: 2025-08-20T06:56:02.889Z
Learnt from: eopXD
Repo: NVIDIA/TensorRT-LLM PR: 6768
File: cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp:577-579
Timestamp: 2025-08-20T06:56:02.889Z
Learning: In cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp, maxSequenceLength is now enforced as a non-optional argument in the BlockManager constructor, so concerns about std::nullopt defaulting to 0 are not applicable. When windowSize > maxSequenceLength, a warning should be added instead of handling optional parameter cases.
Applied to files:
tensorrt_llm/_torch/pyexecutor/executor_request_queue.py
📚 Learning: 2025-08-15T06:46:54.897Z
Learnt from: eopXD
Repo: NVIDIA/TensorRT-LLM PR: 6767
File: cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp:0-0
Timestamp: 2025-08-15T06:46:54.897Z
Learning: In cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp addToken function, newly allocated blocks are unshared by design. The beam search path in addToken (when sequence.getNumTokens() > windowSize) is currently broken/non-functional with SWA, so the block allocation doesn't follow a shared-then-unshared pattern.
Applied to files:
tensorrt_llm/_torch/pyexecutor/resource_manager.py
📚 Learning: 2025-08-21T09:41:49.347Z
Learnt from: eopXD
Repo: NVIDIA/TensorRT-LLM PR: 6768
File: cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp:2010-2045
Timestamp: 2025-08-21T09:41:49.347Z
Learning: In cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp, updateSequenceCacheBlockOffsets is specifically for updating bookkeeping when blocks are added during the context phase, not for refreshing offsets after detach operations. During detach operations, GenerationRequest::removeFrontBlock handles the necessary cache block bookkeeping internally.
Applied to files:
tensorrt_llm/_torch/pyexecutor/resource_manager.py
📚 Learning: 2025-08-20T06:48:45.368Z
Learnt from: eopXD
Repo: NVIDIA/TensorRT-LLM PR: 6768
File: cpp/include/tensorrt_llm/batch_manager/kvCacheManager.h:0-0
Timestamp: 2025-08-20T06:48:45.368Z
Learning: In cpp/tensorrt_llm/batch_manager/kvCacheManager.cpp, updateSequenceCacheBlockOffsets is only called when adding a sequence, not during detach operations. During detach, the cache block bookkeeping is handled by GenerationRequest::removeFrontBlock.
Applied to files:
tensorrt_llm/_torch/pyexecutor/resource_manager.py
🧬 Code graph analysis (1)
tensorrt_llm/_torch/pyexecutor/resource_manager.py (5)
tensorrt_llm/_torch/attention_backend/trtllm.py (1)
tokens_per_block(674-678)tensorrt_llm/runtime/generation.py (1)
tokens_per_block(1218-1219)tensorrt_llm/_torch/distributed/communicator.py (2)
cp_size(56-57)cp_rank(68-69)tensorrt_llm/mapping.py (1)
cp_rank(560-561)tensorrt_llm/_torch/device_mesh.py (1)
cp_rank(84-86)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: Pre-commit Check
🔇 Additional comments (2)
tensorrt_llm/_torch/pyexecutor/model_engine.py (1)
1668-1688: Guard Helix-only fields to avoid potential AttributeError on dummy/warmup requests.The new Helix logic unconditionally accesses
request.py_helix_is_inactive_rankandrequest.seqlen_this_rank_cpwithin theif self.mapping.has_cp_helix()block, which may not hold these attributes for:
- Dummy requests created via
KVCacheManager.add_dummy_requests(if used during warmup)- Any path using Helix mapping without initializing these fields through
_merge_helix_requestsWhile
position_idcomputation is guarded by theif not self.is_warmup and not request.is_cuda_graph_dummycheck, the subsequent attribute accesses are not. Use defensive defaults to gracefully handle missing attributes:if self.mapping.has_cp_helix(): if not self.is_warmup and not request.is_cuda_graph_dummy: position_id = request.total_input_len_cp + request.py_decoding_iter - 1 - if request.py_helix_is_inactive_rank: - past_seen_token_num = request.seqlen_this_rank_cp - else: - # Discount the token added to active rank in resource manager as it hasn't - # been previously seen. - past_seen_token_num = request.seqlen_this_rank_cp - 1 + helix_inactive = getattr(request, "py_helix_is_inactive_rank", False) + seqlen_this_rank_cp = getattr(request, "seqlen_this_rank_cp", + past_seen_token_num) + if helix_inactive: + past_seen_token_num = seqlen_this_rank_cp + else: + # Discount the token added to active rank in resource manager as it hasn't + # been previously seen. + past_seen_token_num = max(seqlen_this_rank_cp - 1, 0)tensorrt_llm/_torch/pyexecutor/resource_manager.py (1)
470-485: Helix decode round-robin implementation is sound, but defensive initialization ofseqlen_this_rank_cpis needed.The new code correctly distributes decode blocks across CP ranks in round-robin fashion and properly skips KV allocation on inactive ranks. However,
req.seqlen_this_rank_cp += 1will raiseAttributeErrorif the attribute is not initialized on all request objects—particularly for dummy/warmup requests created byadd_dummy_requestsor any requests bypassing_merge_helix_requests.Add defensive initialization:
if decode_block_id % self.mapping.cp_size == self.mapping.cp_rank: req.py_helix_is_inactive_rank = False # Initialize per-rank seqlen if not already set (e.g., for dummy/warmup). if not hasattr(req, "seqlen_this_rank_cp"): base = getattr(req, "cached_tokens", getattr(req, "py_prompt_len", 0)) req.seqlen_this_rank_cp = base req.seqlen_this_rank_cp += 1Also verify that
self.tokens_per_blockaligns with the Helixcp_config['tokens_per_block']used during context partitioning; divergence would cause round-robin ownership to misalign with actual sharding.
|
/bot run |
|
/bot run --disable-fail-fast |
|
PR_Github #27159 [ run ] triggered by Bot. Commit: |
|
PR_Github #27160 [ run ] triggered by Bot. Commit: |
|
PR_Github #27159 [ run ] completed with state |
|
PR_Github #27160 [ run ] completed with state |
brb-nv
force-pushed
the
user/brb/load-balance-helix-decode-tokens
branch
from
6b60df1 to
aa61480
Compare
…rallelism Signed-off-by: Balaram Buddharaju <169953907+brb-nv@users.noreply.github.com>
brb-nv
force-pushed
the
user/brb/load-balance-helix-decode-tokens
branch
from
268fe67 to
15a0c2b
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #27553 [ run ] triggered by Bot. Commit: |
|
PR_Github #27553 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #27564 [ run ] triggered by Bot. Commit: |
|
PR_Github #27564 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #27626 [ run ] triggered by Bot. Commit: |
4 participants
Description
Background:
Example:
ISL=400, token_per_block=100, cpSize=4, token range notation is left-inclusive.
Before:
Iteration 1: <0-100> (Rank 0), <100-200> (Rank 1), <200-300> (Rank 2), <300-400> (Rank 3)
Iteration 400: <0-100> (Rank 0), <100-200> (Rank 1), <200-300> (Rank 2), <300-800> (Rank 3)
KV cache is only growing on rank3 during decode. ranks 0,1,2 will have exactly what they received from prefill-decode transmission.
After:
Iteration 1: <0-100> (Rank 0), <100-200> (Rank 1), <200-300> (Rank 2), <300-400> (Rank 3)
Iteration 400: <0-100, 400-500> (Rank 0), <100-200, 500-600> (Rank 1), <200-300, 600-700> (Rank 2), <300-400, 700-800> (Rank 3)
First decode block goes to rank0, second decode block goes to rank1, third decode block goes to rank2, fourth decode block goes to rank3. And then the cycle repeats.
Thanks to @Tabrizian for discussing the example with me.
Test Coverage
Existing test:
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.
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.
Summary by CodeRabbit
✏️ Tip: You can customize this high-level summary in your review settings.