Build

Workflow runs for Build

• Fix THD P2P pad detection and tail-zero gating

  failed:
  Build#20899:Pull request#3506synchronizebyRPalmr
  RPalmr:fix/thd-p2p-pad-detect

  Sep 10, 202618d 18h 19m 55s
  More actions for workflow run 20899

• Fix THD P2P pad detection and tail-zero gating

  failed:
  Build#20898:Pull request#3506openedbyRPalmr
  RPalmr:fix/thd-p2p-pad-detect

  Sep 10, 202618d 18h 21m 7s
  More actions for workflow run 20898

• [Pytorch] MegaMOE Fused Op Integration

  failed:
  Build#20897:Pull request#3350synchronizebyvthumbe1503
  vthumbe1503:dispatch_combine_op

  Sep 10, 20266h 0m 10s
  More actions for workflow run 20897

• Thread tanh logit softcapping through FlashAttention (FA2 and FA3)

  failed:
  Build#20896:Pull request#3391synchronizebynvegesna-netizen
  nvegesna-netizen:nvegesna/gemma2-softcap-core

  Sep 10, 20266h 1m 7s
  More actions for workflow run 20896

• Thread tanh logit softcapping through FlashAttention (FA2 and FA3)

  cancelled:
  Build#20895:Pull request#3391synchronizebycyanguwa
  nvegesna-netizen:nvegesna/gemma2-softcap-core

  Sep 10, 20263h 34m 20s
  More actions for workflow run 20895

• [PyTorch] Support FP8 weight caching in compiled Linear

  failed:
  Build#20894:Pull request#3505openedbypggPL
  pggPL:linear_compile_microbatching

  Sep 10, 20266h 1m 4s
  More actions for workflow run 20894

• [PyTorch] DeepSeekV3Layer: full MoE transformer layer (MLA + DeepSeek MoE)

  failed:
  Build#20893:Pull request#3471synchronizebypggPL
  pggPL:deepseek_v3_layer

  Sep 10, 20266h 0m 39s
  More actions for workflow run 20893

• [PyT] Check FA3 backward compatibility during backend selection

  failed:
  Build#20892:Pull request#3490synchronizebyyuweih205
  yuweih205:fa3-mismatched-dims-training

  Sep 10, 202630d 0h 3m 23s
  More actions for workflow run 20892

• [PyTorch] Support FA3/FA4 with pad_between_seqs=True

  failed:
  Build#20891:Pull request#3205synchronizebyxiaoyao0115
  xiaoyao0115:feat/fa-pad-between-seq

  Sep 10, 20266h 0m 30s
  More actions for workflow run 20891

• [PyTorch] Support FA3/FA4 with pad_between_seqs=True

  cancelled:
  Build#20890:Pull request#3205synchronizebyxiaoyao0115
  xiaoyao0115:feat/fa-pad-between-seq

  Sep 10, 20263m 48s
  More actions for workflow run 20890

• [Pytorch] MegaMOE Fused Op Integration

  failed:
  Build#20889:Pull request#3350synchronizebyvthumbe1503
  vthumbe1503:dispatch_combine_op

  Sep 10, 20266h 0m 20s
  More actions for workflow run 20889

• MOE Sequential Block with Dispatch and Combine as Basic Ops

  failed:
  Build#20888:Pull request#3503synchronizebyvthumbe1503
  vthumbe1503:nccl_ep_ops

  Sep 10, 20266h 5m 38s
  More actions for workflow run 20888

• MOE Sequential Block with Dispatch and Combine as Basic Ops

  cancelled:
  Build#20887:Pull request#3503synchronizebyvthumbe1503
  vthumbe1503:nccl_ep_ops

  Sep 10, 20261h 3m 32s
  More actions for workflow run 20887

• MOE Sequential Block with Dispatch and Combine as Basic Ops

  cancelled:
  Build#20886:Pull request#3503synchronizebyvthumbe1503
  vthumbe1503:nccl_ep_ops

  Sep 10, 20261m 57s
  More actions for workflow run 20886

• [Pytorch][Attention] Fix FP8 THD backward workspace scaling

  failed:
  Build#20885:Pull request#3504openedbysudhakarsingh27
  sudhakarsingh27:sudhakars/fix-fp8-thd-bwd-workspace

  Sep 10, 20266h 0m 11s
  More actions for workflow run 20885

• MOE Sequential Block with Dispatch and Combine as Basic Ops

  cancelled:
  Build#20884:Pull request#3503openedbyvthumbe1503
  vthumbe1503:nccl_ep_ops

  Sep 10, 202640m 32s
  More actions for workflow run 20884

• [Pytorch] Replace TorchScript fallbacks with eager execution

  failed:
  Build#20883:Pull request#3502openedbysudhakarsingh27
  sudhakarsingh27:sudhakars/remove-torchscript

  Sep 10, 20266h 0m 38s
  More actions for workflow run 20883

• [PyTorch] Add timestep-conditioned AdaptiveLayerNorm to op fuser

  failed:
  Build#20882:Pull request#3501openedbyzupengwang
  zupengwang:codex/adaptive-layer-norm

  Sep 10, 202630d 0h 2m 23s
  More actions for workflow run 20882

• [Common] Experimental CuTeDSL MXFP8 backends in C++ via TVM-FFI

  failed:
  Build#20881:Pull request#3137synchronizebykainzhong
  kainzhong:cutedsl_mxfp8_common

  Sep 10, 20266h 0m 13s
  More actions for workflow run 20881

• [Common] Experimental CuTeDSL MXFP8 backends in C++ via TVM-FFI

  cancelled:
  Build#20880:Pull request#3137synchronizebypre-commit-ci[bot]Bot
  kainzhong:cutedsl_mxfp8_common

  Sep 9, 202635m 55s
  More actions for workflow run 20880

• [Common] Experimental CuTeDSL MXFP8 backends in C++ via TVM-FFI

  cancelled:
  Build#20879:Pull request#3137synchronizebykainzhong
  kainzhong:cutedsl_mxfp8_common

  Sep 9, 20261m 29s
  More actions for workflow run 20879

• Thread tanh logit softcapping through FlashAttention (FA2 and FA3)

  failed:
  Build#20878:Pull request#3391synchronizebynvegesna-netizen
  nvegesna-netizen:nvegesna/gemma2-softcap-core

  Sep 9, 20266h 7m 26s
  More actions for workflow run 20878

• [Common] Experimental CuTeDSL MXFP8 backends in C++ via TVM-FFI

  cancelled:
  Build#20877:Pull request#3137synchronizebykainzhong
  kainzhong:cutedsl_mxfp8_common

  Sep 9, 202639m 20s
  More actions for workflow run 20877

• Produce BOLT-compatible libraries

  failed:
  Build#20876:Pull request#3383synchronizebypre-commit-ci[bot]Bot
  fheinecke:fred/bolt-compatible-libs-1

  Sep 9, 20266h 5m 11s
  More actions for workflow run 20876

• Produce BOLT-compatible libraries

  cancelled:
  Build#20875:Pull request#3383synchronizebyfheinecke
  fheinecke:fred/bolt-compatible-libs-1

  Sep 9, 202632m 27s
  More actions for workflow run 20875