Commits

Breadcrumbs

History for

onmain

User selector

All users

All time

Commit history

Commits on Aug 22, 2025

• [Kernel] Add FP8 support with FlashMLA backend (#22668)

  Show description for 19fe1a0
  MatthewBonanni

  authored
  19fe1a0View commit details

  Copy full SHA for 19fe1a0

  View code at this pointBrowse repository at this point

Commits on Aug 21, 2025

• Make sure that vectorize_with_alignment produced vectorized global loads (#23182)

  elvircrn

  authored
  044931fView commit details

  Copy full SHA for 044931f

  View code at this pointBrowse repository at this point

• [Compile] Fix Compile Warning SM100 Cutlass MLA (#23287)

  Show description for f94bf9b
  yewentao256

  authored
  f94bf9bView commit details

  Copy full SHA for f94bf9b

  View code at this pointBrowse repository at this point

• [CPU] Refactor CPU W8A8 scaled_mm (#23071)

  Show description for 7be5d11
  bigPYJ1151

  authored
  7be5d11View commit details

  Copy full SHA for 7be5d11

  View code at this pointBrowse repository at this point

Commits on Aug 20, 2025

• [Kernel/Quant] Remove the original marlin format and qqq (#23204)

  Show description for 0cdbf5e
  mgoin

  authored
  0cdbf5eView commit details

  Copy full SHA for 0cdbf5e

  View code at this pointBrowse repository at this point

• [Kernel] CUTLASS MoE FP8: Integrate cuda moe permute/unpermute (#23045)

  Show description for b17109b
  shixianc

  authored
  b17109bView commit details

  Copy full SHA for b17109b

  View code at this pointBrowse repository at this point

Commits on Aug 18, 2025

• [Bugfix][CI] Machete kernels: deterministic ordering for more cache hits (#23055)

  Show description for b2fd0b8
  andylolu2

  authored
  b2fd0b8View commit details

  Copy full SHA for b2fd0b8

  View code at this pointBrowse repository at this point

Commits on Aug 17, 2025

• [Kernel] Add cuda kernel for gpt_oss activation (#22951)

  Show description for 4d4061b
  jeejeelee

  authored
  4d4061bView commit details

  Copy full SHA for 4d4061b

  View code at this pointBrowse repository at this point

Commits on Aug 16, 2025

• [Kernel/Quant] Remove AQLM (#22943)

  Show description for 4fc722e
  

  mgoin

  and

  yewentao256

  authored
  4fc722eView commit details

  Copy full SHA for 4fc722e

  View code at this pointBrowse repository at this point

Commits on Aug 15, 2025

• [Fix] enable swap_ab for pplx problem size computation (#22991)

  Show description for 7f89ed2
  

  shixianc

  and

  Shixian Cui

  authored
  7f89ed2View commit details

  Copy full SHA for 7f89ed2

  View code at this pointBrowse repository at this point

• [V0 Deprecation] Remove advance_step (#22969)

  Show description for 1c859a1
  WoosukKwon

  authored
  1c859a1View commit details

  Copy full SHA for 1c859a1

  View code at this pointBrowse repository at this point

• Revert "[Kernel] Add cuda kernel for gpt_oss activation" (#22948)

  simon-mo

  authored
  f1f0d2fView commit details

  Copy full SHA for f1f0d2f

  View code at this pointBrowse repository at this point

• [Kernel] Add cuda kernel for gpt_oss activation (#22538)

  Show description for 81f4b96
  jeejeelee

  authored
  81f4b96View commit details

  Copy full SHA for 81f4b96

  View code at this pointBrowse repository at this point

Commits on Aug 14, 2025

• [Kernel] [Quantization] Add MXFP4 and bias support for marlin kernel (#22428)

  Show description for 33c63e9
  

  108 peopleauthored
  33c63e9View commit details

  Copy full SHA for 33c63e9

  View code at this pointBrowse repository at this point

Commits on Aug 13, 2025

• [ROCm][Bugfix] Fix compilation error in topk softmax fused kernel (#22819)

  Show description for c6cd5ca
  kliuae

  authored
  c6cd5caView commit details

  Copy full SHA for c6cd5ca

  View code at this pointBrowse repository at this point

• [Perf] Support topk softmax fused kernel for broader num_experts (#22211)

  Show description for 4c558cf
  

  shixianc

  and

  Shixian Cui

  authored
  4c558cfView commit details

  Copy full SHA for 4c558cf

  View code at this pointBrowse repository at this point

Commits on Aug 9, 2025

• [ROCm][Misc] Rename the context_len to seq_len in ROCm custom paged attention kernel (#22097)

  Show description for b7c0942
  charlifu

  authored
  b7c0942View commit details

  Copy full SHA for b7c0942

  View code at this pointBrowse repository at this point

Commits on Aug 8, 2025

• [Kernel] Add support for block FP8 on SM120 (NVIDIA 5090 and RTX PRO 6000) (#22131)

  Show description for 3303f13
  0xjunhao

  authored
  3303f13View commit details

  Copy full SHA for 3303f13

  View code at this pointBrowse repository at this point

Commits on Aug 7, 2025

• [v1] - Mamba1 Attention Metadata (#21249)

  Show description for 46a1394
  

  Josephasafg

  and

  asafgardin

  authored
  46a1394View commit details

  Copy full SHA for 46a1394

  View code at this pointBrowse repository at this point

Commits on Aug 5, 2025

• [Feature] Non-contiguous Support for FP8 Quantization (#21961)

  Show description for 4771df7
  

  yewentao256

  and

  mgoin

  authored
  4771df7View commit details

  Copy full SHA for 4771df7

  View code at this pointBrowse repository at this point

Commits on Aug 3, 2025

• Support CUTLASS NVFP4 (w4a4) for Blackwell Geforce GPUs (SM120) (#21309)

  Show description for 789562c
  LopezCastroRoberto

  authored
  789562cView commit details

  Copy full SHA for 789562c

  View code at this pointBrowse repository at this point

Commits on Aug 1, 2025

• [Perf] Optimize `reshape_and_cache_flash` CUDA Kernel (#22036)

  Show description for eefbf4a
  yewentao256

  authored
  eefbf4aView commit details

  Copy full SHA for eefbf4a

  View code at this pointBrowse repository at this point

• [Refactor] Fix Compile Warning #1444-D (#21462)

  Show description for a59cd9d
  yewentao256

  authored
  a59cd9dView commit details

  Copy full SHA for a59cd9d

  View code at this pointBrowse repository at this point

Commits on Jul 30, 2025

• [Hardware][CPU] Build fix for ARM without BF16 (#21848)

  Show description for b876860
  ericcurtin

  authored
  b876860View commit details

  Copy full SHA for b876860

  View code at this pointBrowse repository at this point

• [Perf] Using `__nv_fp8_e4m3` instead of `c10::e4m3` for `per_token_group_quant` (#21867)

  Show description for 1b0a155
  yewentao256

  authored
  1b0a155View commit details

  Copy full SHA for 1b0a155

  View code at this pointBrowse repository at this point

• [Docs] Switch to better markdown linting pre-commit hook (#21851)

  Show description for ba5c5e5
  hmellor

  authored
  ba5c5e5View commit details

  Copy full SHA for ba5c5e5

  View code at this pointBrowse repository at this point

Commits on Jul 29, 2025

• [AMD][CI/Build][Bugfix] Guarding CUDA specific functions by ifndef ROCM (#21766)

  Show description for 12a223e
  gshtras

  authored
  12a223eView commit details

  Copy full SHA for 12a223e

  View code at this pointBrowse repository at this point

Commits on Jul 28, 2025

• [Kernel] SM90 CUTLASS FP8 GEMM: add support for swap AB + kernel tuning (#20396)

  Show description for c6c9122
  

  LyrisZhong

  and

  djmmoss

  authored
  c6c9122View commit details

  Copy full SHA for c6c9122

  View code at this pointBrowse repository at this point

Commits on Jul 27, 2025

• Fix CUDA permute/unpermute for use with DeepGemm Moe (#17934)

  Show description for 57c22e5
  CalebDu

  authored
  57c22e5View commit details

  Copy full SHA for 57c22e5

  View code at this pointBrowse repository at this point

Commits on Jul 26, 2025

• [Bugfix][Apple Silicon] fix missing symbols when build from source on Mac with Apple Silicon (#21380)

  Show description for 9094d11
  zhouyeju

  authored
  9094d11View commit details

  Copy full SHA for 9094d11

  View code at this pointBrowse repository at this point

• [Perf] Cuda Kernel for Int8 Per Token Group Quant (#21476)

  Show description for 75d29cf
  yewentao256

  authored
  75d29cfView commit details

  Copy full SHA for 75d29cf

  View code at this pointBrowse repository at this point

Commits on Jul 25, 2025

• [Kernel] Improve machete memory bound perf (#21556)

  Show description for 136d750
  czhu-cohere

  authored
  136d750View commit details

  Copy full SHA for 136d750

  View code at this pointBrowse repository at this point

Commits on Jul 24, 2025

• [Bug] Fix Compressed Tensor NVFP4 `cutlass_fp4_group_mm` illegal memory access (#21465)

  Show description for e8cb0d0
  yewentao256

  authored
  e8cb0d0View commit details

  Copy full SHA for e8cb0d0

  View code at this pointBrowse repository at this point

• [Bugfix][ROCm] Fix for warp_size uses on host (#21205)

  Show description for 90eeea8
  gshtras

  authored
  90eeea8View commit details

  Copy full SHA for 90eeea8

  View code at this pointBrowse repository at this point

Commits on Jul 23, 2025

• [Bugfix][ROCm][Build] Fix build regression on ROCm (#21393)

  Show description for 3ec7170
  gshtras

  authored
  3ec7170View commit details

  Copy full SHA for 3ec7170

  View code at this pointBrowse repository at this point