Hendrik Holtmann c0dfc89485

SM120 / NVFP4: add device guard and runtime SM dispatch to cutlass_scaled_fp4_mm (#29711) ...

Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: mgoin <mgoin64@gmail.com>

2025-12-01 17:24:18 -08:00

..

attention

[Performance][MLA][ROCm] Remove redundant D2D copy in deepseek (#27457)

2025-11-26 12:45:28 +08:00

core

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

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

cpu

[CPU] Update torch 2.9.1 for CPU backend (#29664)

2025-11-28 13:37:54 +00:00

cutlass_extensions

…

mamba/mamba_ssm

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

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

moe

[Kernel][Quantization] add w4a8 support for marlin kernel (#24722)

2025-11-29 07:19:33 -08:00

quantization

SM120 / NVFP4: add device guard and runtime SM dispatch to cutlass_scaled_fp4_mm (#29711)

2025-12-01 17:24:18 -08:00

quickreduce

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

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

rocm

…

sparse/cutlass

…

activation_kernels.cu

…

cache_kernels.cu

[Perf][Deepseek] optimize gather_and_maybe_dequant_cache kernel's perf for extremely long sequence (#28029)

2025-11-24 19:05:46 -07:00

cache.h

[Perf][Deepseek] optimize gather_and_maybe_dequant_cache kernel's perf for extremely long sequence (#28029)

2025-11-24 19:05:46 -07:00

cub_helpers.h

…

cuda_compat.h

…

cuda_utils_kernels.cu

…

cuda_utils.h

…

cuda_view.cu

Simplify from_blob usage in get_cuda_view_from_cpu_tensor (#29027)

2025-11-22 10:35:32 +00:00

cumem_allocator_compat.h

[Core][AMD] Migrate fully transparent sleep mode to ROCm platform (#12695)

2025-11-12 15:24:12 -08:00

cumem_allocator.cpp

[Core][AMD] Migrate fully transparent sleep mode to ROCm platform (#12695)

2025-11-12 15:24:12 -08:00

custom_all_reduce_test.cu

…

custom_all_reduce.cu

…

custom_all_reduce.cuh

…

custom_quickreduce.cu

…

dispatch_utils.h

[kernel][perf] support uncontiguous input for rms_norm kernel (#28103)

2025-11-20 19:39:09 -08:00

fused_qknorm_rope_kernel.cu

[ROCm][Bugfix] Fix compilation errors with fused_qknorm_rope_kernel.cu (#28682)

2025-11-14 11:06:01 -08:00

launch_bounds_utils.h

…

layernorm_kernels.cu

[kernel][perf] support uncontiguous input for rms_norm kernel (#28103)

2025-11-20 19:39:09 -08:00

layernorm_quant_kernels.cu

[Kernel] Optimize rms_norm kernel (#27931)

2025-11-11 18:02:23 +00:00

ops.h

[Performance][MLA][ROCm] Remove redundant D2D copy in deepseek (#27457)

2025-11-26 12:45:28 +08:00

permute_cols.cu

…

pos_encoding_kernels.cu

…

sampler.cu

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

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

torch_bindings.cpp

[Kernel][Quantization] add w4a8 support for marlin kernel (#24722)

2025-11-29 07:19:33 -08:00

type_convert.cuh

[ROCm] [Bugfix] Fix fused_qknorm_rope_kernel rocm compatibility (#28500)

2025-11-12 05:01:14 -08:00