xiao-llm 01413e0cf5

Fp8 paged attention update (#22222) ...

Signed-off-by: Xiao Yu <xiao.yu@amd.com>
Signed-off-by: xiao-llm <xiao.yu.dc@outlook.com>
Co-authored-by: Xiao Yu <xiao.yu@metamaterial.com>
Co-authored-by: Xiao Yu <xiao.yu@amd.com>
Co-authored-by: Bowen Bao <bowenbao@amd.com>

2025-09-15 10:43:26 -04:00

..

attention

[CI Failure] Fix test_flashinfer_cutlass_mxfp4_mxfp8_fused_moe (#24750)

2025-09-13 07:29:19 +00:00

core

[Kernel] [Quantization] Add MXFP4 and bias support for marlin kernel (#22428)

2025-08-14 11:23:22 -07:00

cpu

[Doc]: fix typos in various files (#24726)

2025-09-12 06:43:12 -07:00

cutlass_extensions

[Perf] Use upstream CUTLASS for SM90 Block FP8 kernel (#23280)

2025-09-11 15:43:14 -07:00

mamba/mamba_ssm

[V1][Mamba1] - FP32 SSM Kernel Support (#23506)

2025-09-01 20:53:00 -07:00

moe

[Kernel][Bugfix] Fix grouped topk cu (#24146)

2025-09-04 12:37:37 +08:00

quantization

[Kernels][DP/EP] Optimize Silu Kernel for R1 (#24054)

2025-09-13 00:17:27 -07:00

quickreduce

[Feature] add quick all reduce (#19744)

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

rocm

Fp8 paged attention update (#22222)

2025-09-15 10:43:26 -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

[Feature] Support Decode Context Parallel (DCP) for MLA (#23734)

2025-09-06 13:24:05 +08:00

cache.h

[Feature] Support Decode Context Parallel (DCP) for MLA (#23734)

2025-09-06 13:24:05 +08:00

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

layernorm_kernels.cu

[Model] New model support for Motif-1-Tiny (#23414)

2025-09-10 23:29:40 -07:00

layernorm_quant_kernels.cu

[perf] Add fused MLA QKV + strided layernorm (#21116)

2025-07-22 07:07:44 -07:00

ops.h

[Kernels][DP/EP] Optimize Silu Kernel for R1 (#24054)

2025-09-13 00:17:27 -07:00

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

[BUG] Fix #20484. Support empty sequence in cuda penalty kernel (#20491)

2025-07-05 19:38:02 -07:00

torch_bindings.cpp

[Kernels][DP/EP] Optimize Silu Kernel for R1 (#24054)

2025-09-13 00:17:27 -07:00

type_convert.cuh

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

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