kanshan/TensorRT-LLMs
1
0
Fork 0
mirror of https://github.com/NVIDIA/TensorRT-LLM.git synced 2026-01-14 06:27:45 +08:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity
e5400eeae0
TensorRT-LLMs/cpp/tensorrt_llm/kernels/trtllmGenKernels
History
Dom Brown 44fb3c1673
[TRTLLM-5770] feat: Integrate TRT-LLM Gen FP8 block scale MoE with Pytorch workflow kernel autotuner (#5207) 
- Adds a new Python custom op (fp8_block_scale_moe_runner) and a FP8BlockScaleMoERunner class for autotuning.
- Updates C++ MoE and batched GEMM kernels to accept a configIndex for workspace sizing and execution.
- Extends the unit test to run both autotuned and non-autotuned code paths.

Signed-off-by: Dom Brown <3886319+DomBrown@users.noreply.github.com>
2025-06-17 21:01:56 +08:00
..
batchedGemm [TRTLLM-5770] feat: Integrate TRT-LLM Gen FP8 block scale MoE with Pytorch workflow kernel autotuner (#5207) 2025-06-17 21:01:56 +08:00
blockScaleMoe [TRTLLM-5770] feat: Integrate TRT-LLM Gen FP8 block scale MoE with Pytorch workflow kernel autotuner (#5207) 2025-06-17 21:01:56 +08:00
fmha [https://nvbugspro.nvidia.com/bug/5295470] support headDim 256 for blackwell fmha kernels (#5164) 2025-06-13 23:01:01 +08:00
gemm feat: update DeepSeek FP8 TRT-LLM Gen cubins (#4643) 2025-06-03 14:07:54 -07:00
gemmGatedAct feat: update DeepSeek FP8 TRT-LLM Gen cubins (#4643) 2025-06-03 14:07:54 -07:00
CMakeLists.txt feat: update DeepSeek FP8 TRT-LLM Gen cubins (#4643) 2025-06-03 14:07:54 -07:00