Tyler Michael SmithandClaude Opus 4.6 1e114365c9 Add TrtLLM FP8 support for DeepEP v2 + modular kernel test framework ...

- Register DeepEPV2PrepareAndFinalize and TrtLlmFp8ExpertsModular in
  modular kernel test framework (mk_objects.py)
- Add model_config to test VllmConfig so all2all_utils can read
  enforce_eager for cudagraph detection
- Add weight conversion for TrtLLM BlockMajorK format in test framework
- Separate DeepEP v1/v2 dependency checks in test validation
- Allow TrtLlmFp8ExpertsModular with DeepEP v2 parallel config
- Auto-select TrtLLM backend for DeepEP v2 contiguous layout on Blackwell
- Remove unnecessary torch.where sanitization in DeepEP v2 decode path
- Replace torch.cat with pre-allocated tensor in prefill path
- Always create ExpertTokensMetadata in decode mode

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com>

2026-05-03 16:57:03 -04:00

..

__init__.py

[Misc] Add unit tests for MoE ModularKernel combinations + Profiling utility (#20449)

2025-07-11 07:51:46 -07:00

cli_args.py

[Feature]: Remove Chunking From FusedMoE (#34086)

2026-03-12 14:24:38 -04:00

common.py

Add TrtLLM FP8 support for DeepEP v2 + modular kernel test framework

2026-05-03 16:57:03 -04:00

make_feature_matrix.py

[Feature]: Remove Chunking From FusedMoE (#34086)

2026-03-12 14:24:38 -04:00

mk_objects.py

Add TrtLLM FP8 support for DeepEP v2 + modular kernel test framework

2026-05-03 16:57:03 -04:00

parallel_utils.py

[MoE Refactor] Add more MoE layer tests (#39349)

2026-04-21 18:12:36 -04:00

profile_modular_kernel.py

[Feature]: Remove Chunking From FusedMoE (#34086)

2026-03-12 14:24:38 -04:00