Build

Workflow runs for Build

• feat: vmm slot for local cuda graph offload

  failed:
  Build#20999:Pull request#3544synchronizebypre-commit-ci[bot]Bot
  Xdydy:release_v2.13

  Sep 18, 202630d 0h 2m 44s
  More actions for workflow run 20999

• feat: vmm slot for local cuda graph offload

  failed:
  Build#20998:Pull request#3544openedbyXdydy
  Xdydy:release_v2.13

  Sep 18, 202630d 0h 3m 29s
  More actions for workflow run 20998

• vmm support for cuda graph offload

  failed:
  Build#20997:Pull request#3543synchronizebypre-commit-ci[bot]Bot
  Xdydy:Xdydy/v2.13_vmm

  Sep 18, 202611m 26s
  More actions for workflow run 20997

• vmm support for cuda graph offload

  failed:
  Build#20996:Pull request#3543openedbyXdydy
  Xdydy:Xdydy/v2.13_vmm

  Sep 18, 202613m 33s
  More actions for workflow run 20996

• [PyTorch] Add native transport for dynamic context parallelism

  failed:
  Build#20995:Pull request#3420synchronizebypre-commit-ci[bot]Bot
  xiaoyao0115:cp-p2p-transport

  Sep 17, 20265h 1m 55s
  More actions for workflow run 20995

• [PyTorch] Add native transport for dynamic context parallelism

  cancelled:
  Build#20994:Pull request#3420synchronizebyxiaoyao0115
  xiaoyao0115:cp-p2p-transport

  Sep 17, 202642s
  More actions for workflow run 20994

• [PyTorch] Add native transport for dynamic context parallelism

  cancelled:
  Build#20993:Pull request#3420synchronizebyxiaoyao0115
  xiaoyao0115:cp-p2p-transport

  Sep 17, 202658m 31s
  More actions for workflow run 20993

• [Common] Add non-TMA MXFP8 cast-only kernels for specialized rowwise-only and row+colwise

  failed:
  Build#20992:Pull request#3459synchronizebytdophung
  tdophung:mxfp8-register-cast

  Sep 14, 20266h 5m 11s
  More actions for workflow run 20992

• Support building docs for publishing via GHA

  failed:
  Build#20991:Pull request#3515synchronizebyfheinecke
  fred/support-public-doc-builds-1

  Sep 14, 20266h 0m 34s
  More actions for workflow run 20991

• Support building docs for publishing via GHA

  cancelled:
  Build#20990:Pull request#3515openedbyfheinecke
  fred/support-public-doc-builds-1

  Sep 14, 20262m 50s
  More actions for workflow run 20990

• [PyTorch] torch.compile support for FusedAttention

  failed:
  Build#20989:Pull request#3472synchronizebypggPL
  pggPL:fused_attention_torch_compile

  Sep 14, 20266h 0m 8s
  More actions for workflow run 20989

• [Common] Add non-TMA MXFP8 cast-only kernels for specialized rowwise-only and row+colwise

  cancelled:
  Build#20988:Pull request#3459synchronizebytdophung
  tdophung:mxfp8-register-cast

  Sep 14, 202627m 12s
  More actions for workflow run 20988

• [JAX] Add EP borrowed-comm path via XLA FFI collectives extension

  failed:
  Build#20987:Pull request#3452synchronizebyphu0ngng
  phu0ngng:ep_jax_single_process

  Sep 14, 20266h 0m 41s
  More actions for workflow run 20987

• [PyTorch] Support dynamic recipe updates during training

  failed:
  Build#20986:Pull request#3343synchronizebynegvet
  negvet:mid_train_recipe_update

  Sep 14, 20266h 1m 28s
  More actions for workflow run 20986

• [PyTorch Debug] Add scale_inv_std stat and skip NVFP4 layers in LogFp8TensorStats

  failed:
  Build#20985:Pull request#3044synchronizebypggPL
  pggPL:debug_log_fp_fixes

  Sep 14, 20266h 0m 41s
  More actions for workflow run 20985

• [PyTorch] Split FusedAttnFunc into single-argument forward/backward helpers

  failed:
  Build#20984:Pull request#3480synchronizebypggPL
  pggPL:fused_attn_func_refactor

  Sep 14, 20266h 1m 21s
  More actions for workflow run 20984

• [PyTorch] Split FusedAttnFunc into single-argument forward/backward helpers

  cancelled:
  Build#20983:Pull request#3480synchronizebypggPL
  pggPL:fused_attn_func_refactor

  Sep 14, 202657m 42s
  More actions for workflow run 20983

• [PyTorch] Support dynamic recipe updates during training

  cancelled:
  Build#20982:Pull request#3343synchronizebynegvet
  negvet:mid_train_recipe_update

  Sep 14, 20265h 45m 46s
  More actions for workflow run 20982

• [PyTorch] Support dynamic recipe updates during training

  cancelled:
  Build#20981:Pull request#3343synchronizebynegvet
  negvet:mid_train_recipe_update

  Sep 14, 202657s
  More actions for workflow run 20981

• Add MXFP8 support with cuBLASMp

  failed:
  Build#20980:Pull request#3145synchronizebyalmogsegal
  almogsegal:add-mxfp8-and-nvfp4-with-cublasmp

  Sep 14, 20266h 0m 27s
  More actions for workflow run 20980

• Abstract CUDA hardcodes into configurable te_device_type / te_platform

  failed:
  Build#20979:Pull request#3513synchronizebylxd-cumt
  lxd-cumt:cuda_patch_main_latest

  Sep 14, 202630d 0h 5m 1s
  More actions for workflow run 20979

• Abstract CUDA hardcodes into configurable te_device_type / te_platform

  failed:
  Build#20978:Pull request#3513synchronizebylxd-cumt
  lxd-cumt:cuda_patch_main_latest

  Sep 14, 202630d 0h 2m 55s
  More actions for workflow run 20978

• Abstract CUDA hardcodes into configurable te_device_type / te_platform

  failed:
  Build#20977:Pull request#3513synchronizebylxd-cumt
  lxd-cumt:cuda_patch_main_latest

  Sep 14, 202630d 0h 3m 18s
  More actions for workflow run 20977

• Abstract CUDA hardcodes into configurable te_device_type / te_platform

  failed:
  Build#20976:Pull request#3513synchronizebylxd-cumt
  lxd-cumt:cuda_patch_main_latest

  Sep 14, 202630d 0h 0m 28s
  More actions for workflow run 20976

• Abstract CUDA hardcodes into configurable te_device_type / te_platform

  failed:
  Build#20975:Pull request#3513openedbylxd-cumt
  lxd-cumt:cuda_patch_main_latest

  Sep 14, 202630d 0h 0m 18s
  More actions for workflow run 20975