3 people 294c805f1d Early exit for MoE LoRA kernels (#27131) ...

Signed-off-by: gnovack <gnovack@amazon.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>

2025-11-03 20:22:17 +08:00

..

attention

[Attention] Tune CUTLASS MLA num_splits (#26846)

2025-10-16 06:36:09 -07:00

core

[small][batch invariance] Rename the env and internal flags to simplify usage (#26855)

2025-10-16 21:40:25 +00:00

cpu

[cpu][fix] Fix onednn_mm crash on consecutive matmuls with same M,K,N and different dtype (#27472)

2025-10-24 15:57:48 +00:00

cutlass_extensions

Update Optional[x] -> x | None and Union[x, y] to x | y (#26633)

2025-10-12 09:51:31 -07:00

mamba/mamba_ssm

[V1] [Hybrid] Mamba1 Automatic Prefix Caching (#26377)

2025-11-02 04:16:23 -08:00

moe

Early exit for MoE LoRA kernels (#27131)

2025-11-03 20:22:17 +08:00

quantization

[Kernel] Add GPTQv2 format support for low-bit or asymmetric quantization, by adapting gptq_gemm (#26092)

2025-10-23 23:26:13 -04:00

quickreduce

[Bugfix][Rocm] fix qr error when different inp shape (#25892)

2025-10-13 10:04:21 -07:00

rocm

[Refactor] Refactor FP8 & INT8 Quant Folder inside w8a8 (#25293)

2025-10-08 10:20:48 -04:00

sparse/cutlass

[feat]: CUTLASS block scaled group gemm for SM100 (#19757)

2025-07-04 12:58:04 -06:00

activation_kernels.cu

[Kernel] Add cuda kernel for gpt_oss activation (#22951)

2025-08-17 05:03:24 +00:00

cache_kernels.cu

[Refactor] Refactor FP8 & INT8 Quant Folder inside w8a8 (#25293)

2025-10-08 10:20:48 -04:00

cache.h

Add gather_indexer_k_quant_cache kernel (#25931)

2025-10-08 04:58:57 +00:00

cub_helpers.h

[Refactor] Refactor FP8 & INT8 Quant Folder inside w8a8 (#25293)

2025-10-08 10:20:48 -04:00

cuda_compat.h

[Bugfix][ROCm] Fix for warp_size uses on host (#21205)

2025-07-24 00:37:19 -07:00

cuda_utils_kernels.cu

[NVIDIA] Support nvfp4 quantization (#12784)

2025-02-12 19:51:51 -08:00

cuda_utils.h

[Attention] MLA with chunked prefill (#12639)

2025-02-21 15:30:12 -08:00

cuda_view.cu

[V1] Fully Transparent Implementation of CPU Offloading (#15354)

2025-03-31 20:22:34 +08:00

cumem_allocator.cpp

[core] improve error handling when wake up from sleep mode (#12981)

2025-02-10 09:38:57 +08:00

custom_all_reduce_test.cu

[Distributed] Add custom allreduce support for ROCM (#14125)

2025-03-31 22:49:12 -07:00

custom_all_reduce.cu

[Distributed] Add custom allreduce support for ROCM (#14125)

2025-03-31 22:49:12 -07:00

custom_all_reduce.cuh

[Kernels] Enable Torch Symmetric Memory All-Reduce By Default (#24111)

2025-09-11 09:45:31 -07:00

custom_quickreduce.cu

[Feature] add quick all reduce (#19744)

2025-06-26 20:54:24 -07:00

dispatch_utils.h

[Bugfix][Misc] Fix silu_and_mul_nvfp4_quant issue and extract common utils for nvfp4 kernel source files (#23727)

2025-09-04 14:25:45 -07:00

launch_bounds_utils.h

Update launch_bounds_utils.h for correct compile on Multiple Cuda Arch - PTXAS out of range Warning (#25843)

2025-09-30 19:18:19 -07:00

layernorm_kernels.cu

[Chore] Remove unused PolyNorm layer (#27110)

2025-10-17 19:03:43 +00:00

layernorm_quant_kernels.cu

[torch.compile] Enable attention and allreduce fusion without custom ops enabled (#24604)

2025-10-17 08:10:23 -06:00

ops.h

[V1] [Hybrid] Mamba1 Automatic Prefix Caching (#26377)

2025-11-02 04:16:23 -08:00

permute_cols.cu

[Kernel] (2/N) Machete - Integrate into CompressedTensorsWNA16 and GPTQMarlin (#7701)

2024-09-23 13:46:26 -04:00

pos_encoding_kernels.cu

[Chore] Remove unused batched RoPE op & kernel (#24789)

2025-09-13 00:08:20 -07:00

sampler.cu

[Deepseek v3.2] Remove extra logics in indexer (#26465)

2025-10-21 23:34:03 +00:00

torch_bindings.cpp

[V1] [Hybrid] Mamba1 Automatic Prefix Caching (#26377)

2025-11-02 04:16:23 -08:00

type_convert.cuh

[torch.compile] Fuse RMSNorm with quant (#9138)

2024-11-08 21:20:08 +00:00