Commits

History for

Commits on Jun 4, 2025

[KERNEL] Sampler. CUDA kernel for applying repetition penalty (#18437)
vadiklyutiy
authored

Commits on Jun 3, 2025

[ROCm][Build] Clean up the ROCm build (#19040)
gshtras
authored

Commits on May 28, 2025

[Build] Fixes for CMake install (#18570)
ProExpertProg
authored

Commits on May 27, 2025

feat(rocm-support): support mamba2 on rocm (#18565)
almersawi
authored

Commits on May 23, 2025

[Bugfix][Build/CI] Fixup CUDA compiler version check for CUDA_SUPPORTED_ARCHS (#18579)
simon-mo
authored

Commits on May 22, 2025

[Build/CI] Fix CUDA 11.8 build (#17679)

tlrmchlsmth
and
LucasWilkinson
authored

Commits on May 15, 2025

[Build] Allow shipping PTX on a per-file basis (#18155)
LucasWilkinson
authored

Commits on May 13, 2025

Implements dual-chunk-flash-attn backend for dual chunk attention with sparse attention support (#11844)
sighingnow
authored

Commits on May 9, 2025

[Hardware/NVIDIA/Kernel] Enable nvidia/DeepSeek-R1-FP4 Model (#16362)
pavanimajety
authored

Commits on May 8, 2025

Add cutlass support for blackwell fp8 blockwise gemm (#14383)
wenscarl
authored

Commits on May 5, 2025

Commits on May 3, 2025

[Build/CI] Upgrade CUTLASS to 3.9.1 (#17602)
tlrmchlsmth
authored

Commits on May 2, 2025

permute/unpermute kernel for moe optimization (#14568)
CalebDu
authored

Commits on May 1, 2025

[torch.compile] Add torch inductor pass for fusing silu_and_mul with subsequent scaled_fp8_quant operations (#10867)
SageMoore
authored

Commits on Apr 30, 2025

Update PyTorch to 2.7.0 (#16859)
huydhn
authored

Commits on Apr 27, 2025

[NVIDIA] Support Cutlass MLA for Blackwell GPUs (#16032)
kaixih
authored

Commits on Apr 22, 2025

[Performance][ROCm] Add skinny gemms for unquantized linear on ROCm (#15830)

charlifu
and
tlrmchlsmth
authored

Commits on Apr 15, 2025

[Kernel] moe wna16 marlin kernel (#14447)

jinzhen-lin
and
mgoin
authored

Commits on Apr 11, 2025

[Kernel] support merge_attn_states CUDA kernel, 3x speedup (#16173)
DefTruth
authored

Commits on Apr 1, 2025

[Distributed] Add custom allreduce support for ROCM (#14125)

ilmarkov
and
ilmarkov
authored

Commits on Mar 31, 2025

Commits on Mar 29, 2025

[ROCm][AMD][Build] Update AMD supported arch list (#15632)
gshtras
authored

Commits on Mar 27, 2025

[Kernel] CUTLASS grouped gemm fp8 MoE kernel (#13972)

ElizaWszola
and
LucasWilkinson
authored

Commits on Mar 14, 2025

Commits on Mar 12, 2025

Commits on Mar 11, 2025

Commits on Mar 8, 2025

[Build/BugFix] Fix hopper 12.8 build (#14354)

LucasWilkinson
and
tlrmchlsmth
authored

Commits on Mar 5, 2025

Disable GPTQ AllSpark kernels for CUDA Compiler < 12.0 (#14157)
mgoin
authored

Commits on Mar 4, 2025

add cutlass support for blackwell fp8 gemm (#13798)
kushanam
authored

Commits on Mar 1, 2025

[Misc][Kernel]: Add GPTQAllSpark Quantization (#12931)
wyajieha
authored