No GPU

Workflow runs for No GPU

• docs: install Xcode's Metal Toolchain component before a Mac build

  currently running:
  No GPU#989:Pull request#1049openedbyakol1
  akol1:docs/metal-toolchain-component

  Oct 18, 2026In progress
  More actions for workflow run 989

• DeepSeek V4.1: the lane backend — one stream per slot on the lane core

  failed:
  No GPU#988:Pull request#827synchronizebyakol1
  akol1:pr-dsv41-d-lane

  Oct 18, 20268m 12s
  More actions for workflow run 988

• DeepSeek V4.1: the lane backend — one stream per slot on the lane core

  failed:
  No GPU#987:Pull request#827synchronizebyakol1
  akol1:pr-dsv41-d-lane

  Oct 18, 20268m 13s
  More actions for workflow run 987

• DeepSeek V4.1: the fp64 reference tests — every operator against a three-way harness

  failed:
  No GPU#986:Pull request#826synchronizebyakol1
  akol1:pr-dsv41-c-tests

  Oct 18, 20268m 35s
  More actions for workflow run 986

• DeepSeek V4.1: the family — configuration, Engram tables, resident weights and the Metal operators

  failed:
  No GPU#985:Pull request#825synchronizebyakol1
  akol1:pr-dsv41-bc-family

  Oct 18, 20269m 44s
  More actions for workflow run 985

• DeepSeek V4.1: the Engram token map built in Zig, byte-identical to the release's

  failed:
  No GPU#984:Pull request#824synchronizebyakol1
  akol1:pr-dsv41-a-token-map

  Oct 18, 20268m 32s
  More actions for workflow run 984

• Nemotron 3.5 Bolt (nemotron_h_puzzle) on Metal — serves end to end

  requires action with the application:
  No GPU#983:Pull request#1038synchronizebytrickyfalcon
  bytebunkerlabs:nemotron-h-puzzle

  Oct 18, 2026Action required
  More actions for workflow run 983

• GLM-5.3-Flash on two CUDA ranks: prompts admitted together (prefill_many)

  failed:
  No GPU#982:Pull request#1014synchronizebyBobClawblaw
  BobClawblaw:pr/glm-prefill-many

  Oct 18, 20265m 58s
  More actions for workflow run 982

• Nemotron 3.5 Bolt (nemotron_h_puzzle) on Metal — serves end to end

  requires action with the application:
  No GPU#981:Pull request#1038synchronizebytrickyfalcon
  bytebunkerlabs:nemotron-h-puzzle

  Oct 18, 2026Action required
  More actions for workflow run 981

• qwen3_5 CUDA: a captured bucket's padding rows no longer go through the routed experts (35B, every multi-row forward)

  failed:
  No GPU#980:Pull request#1047openedbyjschmied
  jschmied:q35-moe-live-rows

  Oct 18, 20268m 36s
  More actions for workflow run 980

• Flash Next CUDA: prompt processing 1.4-1.8x faster on GB10, bit for bit with 1.1.0

  failed:
  No GPU#979:Pull request#1046openedbyMirrdhyn
  hygurlabs:gb10-110c

  Oct 18, 20268m 16s
  More actions for workflow run 979

• Windows: main's host tests build again on Windows hosts

  completed successfully:
  No GPU#978:Pull request#1034synchronizebyjayleaton
  jayleaton:fix/windows-host-build

  Oct 18, 202611m 27s
  More actions for workflow run 978

• Nemotron 3.5 Bolt (nemotron_h_puzzle) on Metal — serves end to end

  requires action with the application:
  No GPU#977:Pull request#1038synchronizebytrickyfalcon
  bytebunkerlabs:nemotron-h-puzzle

  Oct 18, 2026Action required
  More actions for workflow run 977

• Benchmark: keep each attempt's draft counts beside the receipt

  failed:
  No GPU#976:Pull request#1045synchronizebyThinkOffApp
  ThinkOffApp:benchmark-drafter-counts

  Oct 18, 20268m 28s
  More actions for workflow run 976

• HIP: refuse --drafter until the engine runs DFlash2

  failed:
  No GPU#975:Pull request#1044synchronizebyThinkOffApp
  ThinkOffApp:hip-refuses-drafter

  Oct 18, 20268m 40s
  More actions for workflow run 975

• Nemotron 3.5 Bolt (nemotron_h_puzzle) on Metal — serves end to end

  requires action with the application:
  No GPU#974:Pull request#1038synchronizebytrickyfalcon
  bytebunkerlabs:nemotron-h-puzzle

  Oct 18, 2026Action required
  More actions for workflow run 974

• Benchmark: ROCm receipts name their GPU, CPU and checkpoint the way the board checks them

  failed:
  No GPU#973:Pull request#1043synchronizebyThinkOffApp
  ThinkOffApp:benchmark-rocm-host-identity

  Oct 18, 20268m 19s
  More actions for workflow run 973

• Benchmark: keep each attempt's draft counts beside the receipt

  failed:
  No GPU#972:Pull request#1045openedbyThinkOffApp
  ThinkOffApp:benchmark-drafter-counts

  Oct 18, 20268m 32s
  More actions for workflow run 972

• HIP: refuse --drafter until the engine runs DFlash2

  failed:
  No GPU#971:Pull request#1044openedbyThinkOffApp
  ThinkOffApp:hip-refuses-drafter

  Oct 18, 20268m 28s
  More actions for workflow run 971

• Report unavailable engine health and preserve admission HTTP 503

  failed:
  No GPU#970:Pull request#912synchronizebyakol1
  akol1:codex/engine-unavailable-health

  Oct 18, 20268m 57s
  More actions for workflow run 970

• Benchmark: ROCm receipts name their GPU, CPU and checkpoint the way the board checks them

  failed:
  No GPU#969:Pull request#1043openedbyThinkOffApp
  ThinkOffApp:benchmark-rocm-host-identity

  Oct 18, 20267m 31s
  More actions for workflow run 969

• docs: measure the ModelOpt checkpoint on one GB10

  failed:
  No GPU#968:Pull request#1042openedbyakol1
  akol1:docs/flash-next-modelopt-gb10

  Oct 18, 20268m 22s
  More actions for workflow run 968

• Nemotron 3.5 Bolt (nemotron_h_puzzle) on Metal — serves end to end

  requires action with the application:
  No GPU#967:Pull request#1038synchronizebytrickyfalcon
  bytebunkerlabs:nemotron-h-puzzle

  Oct 18, 2026Action required
  More actions for workflow run 967

• Nemotron 3.5 Bolt (nemotron_h_puzzle) on Metal — serves end to end

  requires action with the application:
  No GPU#966:Pull request#1038synchronizebytrickyfalcon
  bytebunkerlabs:nemotron-h-puzzle

  Oct 18, 2026Action required
  More actions for workflow run 966

• Nemotron 3.5 Bolt (nemotron_h_puzzle) on Metal — serves end to end

  requires action with the application:
  No GPU#965:Pull request#1038synchronizebytrickyfalcon
  bytebunkerlabs:nemotron-h-puzzle

  Oct 18, 2026Action required
  More actions for workflow run 965