1Cat-vLLM

attention

[6/n] Migrate activation kernels, gptq, gguf, non cutlass w8a8 to libtorch stable ABI (continued) (#42663 )

2026-05-20 00:18:12 -07:00

core

[5/n] Migrate CUTLASS MLA, hadamard, awq, allspark and DSV3 fused a gemm to torch stable ABI (continued) (#42339 )

2026-05-13 07:24:39 +00:00

cpu

[CPU][RISC-V] Add VLEN=256 support to RVV attention kernels (#42943 )

2026-05-21 04:50:49 -07:00

cutlass_extensions

[6/n] Migrate activation kernels, gptq, gguf, non cutlass w8a8 to libtorch stable ABI (continued) (#42663 )

2026-05-20 00:18:12 -07:00

libtorch_stable

[6/n] Migrate activation kernels, gptq, gguf, non cutlass w8a8 to libtorch stable ABI (continued) (#42663 )

2026-05-20 00:18:12 -07:00

mamba/mamba_ssm

…

moe

[Refactor] Remove dead cuda kernels (#42767 )

2026-05-18 11:14:21 -07:00

quantization

[6/n] Migrate activation kernels, gptq, gguf, non cutlass w8a8 to libtorch stable ABI (continued) (#42663 )

2026-05-20 00:18:12 -07:00

quickreduce

…

rocm

[ROCm] Add gfx1102/gfx1103 support (#40037 )

2026-04-23 01:32:04 -07:00

async_util.cuh

fused qknorm+rope kernel optimization for SM9.0 (#37376 )

2026-04-12 19:58:37 -07:00

cache_kernels_fused.cu

[Performance][DSR1]: Fused RoPE+KVCache+q_concat for MLA (#40392 )

2026-05-11 14:10:50 +00:00

cache_kernels.cu

Use CU_MEMCPY_SRC_ACCESS_ORDER_ANY for batch KV cache swaps (#39306 )

2026-05-10 05:57:09 +03:00

cache.h

Use CU_MEMCPY_SRC_ACCESS_ORDER_ANY for batch KV cache swaps (#39306 )

2026-05-10 05:57:09 +03:00

concat_mla_q.cuh

…

cub_helpers.h

…

cuda_compat.h

…

cuda_utils_kernels.cu

…

cuda_utils.h

…

cuda_vec_utils.cuh

[6/n] Migrate activation kernels, gptq, gguf, non cutlass w8a8 to libtorch stable ABI (continued) (#42663 )

2026-05-20 00:18:12 -07:00

cuda_view.cu

…

cumem_allocator_compat.h

…

cumem_allocator.cpp

fix(rocm): remove workaround causing invalid argument on Qwen3.5 with TP=2 (#40686 )

2026-05-06 01:38:32 -07:00

custom_all_reduce_test.cu

…

custom_all_reduce.cu

…

custom_all_reduce.cuh

…

custom_quickreduce.cu

…

dispatch_utils.h

…

fused_deepseek_v4_qnorm_rope_kv_insert_kernel.cu

DSv4 fused Q-norm kernel grid refactor (#42353 )

2026-05-22 15:21:33 -07:00

fused_qknorm_rope_kernel.cu

fused qknorm+rope kernel optimization for SM9.0 (#37376 )

2026-04-12 19:58:37 -07:00

launch_bounds_utils.h

…

layernorm_kernels.cu

[Feat] DeepSeek V4 Rebased (#40860 )

2026-04-26 18:31:08 -07:00

layernorm_quant_kernels.cu

[Feat] DeepSeek V4 Rebased (#40860 )

2026-04-26 18:31:08 -07:00

minimax_reduce_rms_kernel.cu

…

minimax_reduce_rms_kernel.h

…

nvfp4_kv_cache_kernels.cu

Add nvfp4 support to reshape_and_cache_flash (#37332 )

2026-04-17 07:28:00 -07:00

ops.h

[6/n] Migrate activation kernels, gptq, gguf, non cutlass w8a8 to libtorch stable ABI (continued) (#42663 )

2026-05-20 00:18:12 -07:00

persistent_topk.cuh

[Bugfix] Fix persistent_topk inter-CTA init race on RadixRowState (#41444 )

2026-05-01 14:42:52 -07:00

pos_encoding_kernels.cu

[Bugfix] Fix rope (#41113 )

2026-04-28 22:42:35 -07:00

sampler.cu

[Feat] DeepSeek V4 Rebased (#40860 )

2026-04-26 18:31:08 -07:00

spinloop.cpp

Add VLLM_USE_SPINLOOP_EXT to use more efficient busy polling (#36517 )

2026-05-11 16:11:49 -07:00

topk.cu

[Bugfix] Fix DeepSeek v4 topk numerical issue for unaligned max-model-len (#42169 )

2026-05-09 20:30:08 -07:00

torch_bindings.cpp

[6/n] Migrate activation kernels, gptq, gguf, non cutlass w8a8 to libtorch stable ABI (continued) (#42663 )

2026-05-20 00:18:12 -07:00

torch_utils.h

[6/n] Migrate activation kernels, gptq, gguf, non cutlass w8a8 to libtorch stable ABI (continued) (#42663 )

2026-05-20 00:18:12 -07:00

type_convert.cuh

…