4 people 08a65e287e [Bugfix][Kernel] hadacore: don't load past the input on a partial last warp (#61944) ...

Signed-off-by: khluu <khluu000@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
Co-authored-by: vllm-agent <claw@inferact.ai>

2026-10-17 00:52:17 -07:00

..

awq

[Misc] Consolidate csrc shared headers and prune dead code (#60721)

2026-10-12 09:46:22 -07:00

cutlass_w4a8

[Misc] Consolidate csrc shared headers and prune dead code (#60721)

2026-10-12 09:46:22 -07:00

fp4

[Misc] Consolidate csrc shared headers and prune dead code (#60721)

2026-10-12 09:46:22 -07:00

fused_kernels

[Perf] per_token_group_quant_8bit: multiply by reciprocal instead of dividing per element (#61343)

2026-10-13 21:37:10 +08:00

gptq

[Misc] Consolidate csrc shared headers and prune dead code (#60721)

2026-10-12 09:46:22 -07:00

hadamard/hadacore

[Bugfix][Kernel] hadacore: don't load past the input on a partial last warp (#61944)

2026-10-17 00:52:17 -07:00

machete

[Misc] Consolidate csrc shared headers and prune dead code (#60721)

2026-10-12 09:46:22 -07:00

marlin

[Misc] Consolidate csrc shared headers and prune dead code (#60721)

2026-10-12 09:46:22 -07:00

w8a8

[GLM 5.3 Perf] Handle two elements at the same time for per token group quant, 2.8% E2E ITL improvement (#60927)

2026-10-13 13:06:41 -07:00

activation_kernels.cu

[Bugfix][Kernel] silu_and_mul_quant: run the vectorized loop (#61716)

2026-10-15 18:16:53 -07:00

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

vectorization_utils.cuh

[Misc] Consolidate csrc shared headers and prune dead code (#60721)

2026-10-12 09:46:22 -07:00

vectorization.cuh

[Misc] Consolidate csrc shared headers and prune dead code (#60721)

2026-10-12 09:46:22 -07:00