Hashem Hashemi 721ae79f50 Improvements to wvSplitKrc skinny GEMM solution (#34304) ...

Signed-off-by: Hashem Hashemi <hashem.hashemi@amd.com>

2026-03-10 09:14:27 -07:00

..

attention

…

core

…

cpu

[Bugfix][CI] fix typos (#34934)

2026-03-05 17:05:46 +00:00

cutlass_extensions

[Kernel] Add enable_sm120_or_later for SM121 (DGX Spark) CUTLASS support (#33517)

2026-02-06 20:28:01 -08:00

mamba/mamba_ssm

[Mamba1] - Kernel Level Chunk Alignment for Prefix Caching (#34798)

2026-03-01 20:40:23 +08:00

moe

[Bugfix][CI] fix typos (#34934)

2026-03-05 17:05:46 +00:00

quantization

[Bugfix][CI] fix typos (#34934)

2026-03-05 17:05:46 +00:00

quickreduce

…

rocm

Improvements to wvSplitKrc skinny GEMM solution (#34304)

2026-03-10 09:14:27 -07:00

sparse/cutlass

[Bugfix] Fix Sparse24 Compressed Tensors models (#33446)

2026-02-11 23:15:16 -08:00

activation_kernels.cu

[Refactor][Kernel] Add global helper to deduplicate vectorized memory ops (#35105)

2026-02-27 16:28:17 -08:00

cache_kernels_fused.cu

Fuse RoPE and MLA KV-cache write (#25774)

2026-01-09 19:18:37 -08:00

cache_kernels.cu

[Attention][Perf][Kernel] Replace torch.cat with vectorized CUDA kernel MLA query concat - DeepSeek-V3.2 (#34917)

2026-03-09 09:50:36 -07:00

cache.h

[Attention][Perf][Kernel] Replace torch.cat with vectorized CUDA kernel MLA query concat - DeepSeek-V3.2 (#34917)

2026-03-09 09:50:36 -07:00

concat_mla_q.cuh

[Attention][Perf][Kernel] Replace torch.cat with vectorized CUDA kernel MLA query concat - DeepSeek-V3.2 (#34917)

2026-03-09 09:50:36 -07:00

cub_helpers.h

…

cuda_compat.h

…

cuda_utils_kernels.cu

…

cuda_utils.h

…

cuda_vec_utils.cuh

[Attention][Perf][Kernel] Replace torch.cat with vectorized CUDA kernel MLA query concat - DeepSeek-V3.2 (#34917)

2026-03-09 09:50:36 -07:00

cuda_view.cu

[Feature] Support CPU Offloading without Pytorch Pinned Memory that leads to doubled allocation (#32993)

2026-02-13 08:11:26 -08:00

cumem_allocator_compat.h

…

cumem_allocator.cpp

[Feature][Core] Support Fabric detection to adapt the MNNVL protocol for the GB series (#33540)

2026-02-02 22:55:46 +08:00

custom_all_reduce_test.cu

…

custom_all_reduce.cu

…

custom_all_reduce.cuh

…

custom_quickreduce.cu

…

dispatch_utils.h

…

dsv3_fused_a_gemm.cu

[Bugfix] Fix DSV3 kernels breaking _C and _moe_C on unsupported arches (#35123)

2026-02-23 17:11:27 -08:00

fused_qknorm_rope_kernel.cu

…

launch_bounds_utils.h

…

layernorm_kernels.cu

…

layernorm_quant_kernels.cu

…

ops.h

[Mamba1] - Kernel Level Chunk Alignment for Prefix Caching (#34798)

2026-03-01 20:40:23 +08:00

permute_cols.cu

…

pos_encoding_kernels.cu

…

sampler.cu

[Perf][Kernel] Add faster topKperRow decode kernel for DeepSeek-V3.2 sparse attention (#33680)

2026-02-10 10:29:52 -05:00

topk.cu

[Bugfix] Replace c10::optional with std::optional in topk kernel (#34467)

2026-02-13 08:30:23 -08:00

torch_bindings.cpp

[Attention][Perf][Kernel] Replace torch.cat with vectorized CUDA kernel MLA query concat - DeepSeek-V3.2 (#34917)

2026-03-09 09:50:36 -07:00

type_convert.cuh

…