Skip to content

Actions: pytorch/FBGEMM

Actions

Build FBGEMM GenAI Aarch64 Linux Wheels

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
build_wheels_genai_linux_aarch64.yml
2,500+ workflow runs
2,500+ workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
2026-08-05 nightly release (d3d1883ae278e28f24719ff7292cfb628359b0d2) Build FBGEMM GenAI Aarch64 Linux Wheels #10888: Commit 294b0c5 pushed by pytorchbot
In progress nightly
nightly
In progress
[ROCm] Add fp8 corner case unit test Build FBGEMM GenAI Aarch64 Linux Wheels #10887: Pull request #6115 opened by avbokovoy
Action required ROCm:abokovoi/add-fp8fnuz-vs-ocp-test
ROCm:abokovoi/add-fp8fnuz-vs-ocp-test
Action required
Fix backward regression in PR #5712 Build FBGEMM GenAI Aarch64 Linux Wheels #10886: Pull request #5825 synchronize by avbokovoy
Action required ROCm:abokovoi/pr_5712_backward_regression
ROCm:abokovoi/pr_5712_backward_regression
Action required
num_stages=0 -> 1 (pytorch/FBGEMM) (#6111) Build FBGEMM GenAI Aarch64 Linux Wheels #10885: Commit d3d1883 pushed by meta-codesync Bot
2h 2m 59s main
main
2h 2m 59s
Add mixed precision support to the HIP optimized warp per row TBE backward kernel (#6020) Build FBGEMM GenAI Aarch64 Linux Wheels #10884: Pull request #6113 opened by q10
2h 3m 15s q10:export-D112378971
q10:export-D112378971
2h 3m 15s
Backout "Revert D95859348" (#6015) Build FBGEMM GenAI Aarch64 Linux Wheels #10883: Pull request #6015 synchronize by pklund
2h 8m 41s pklund:export-D110939120
pklund:export-D110939120
2h 8m 41s
Add grid-stride + ROCm cap to fp8_rowwise quantize kernels (#6110) Build FBGEMM GenAI Aarch64 Linux Wheels #10882: Commit 185a3e7 pushed by meta-codesync Bot
2h 16m 47s main
main
2h 16m 47s
Add grid-stride + ROCm cap to padded_fp8_rowwise quantize kernels Build FBGEMM GenAI Aarch64 Linux Wheels #10881: Pull request #6112 opened by q10
2h 14m 50s q10:export-D113375952
q10:export-D113375952
2h 14m 50s
Add grid-stride + ROCm cap to fused_8bit_rowwise quantize kernels (#6… Build FBGEMM GenAI Aarch64 Linux Wheels #10880: Commit 6c3624d pushed by meta-codesync Bot
1h 6m 58s main
main
1h 6m 58s
num_stages=0 -> 1 (pytorch/FBGEMM) Build FBGEMM GenAI Aarch64 Linux Wheels #10879: Pull request #6111 opened by njriasan
2h 7m 4s njriasan:export-D114753270
njriasan:export-D114753270
2h 7m 4s
Add grid-stride + ROCm cap to fp8_rowwise quantize kernels Build FBGEMM GenAI Aarch64 Linux Wheels #10878: Pull request #6110 opened by q10
2h 6m 18s q10:export-D113375950
q10:export-D113375950
2h 6m 18s
Reland: arch-aware FP8 selection via single-instantiation dispatch Build FBGEMM GenAI Aarch64 Linux Wheels #10877: Pull request #6104 synchronize by q10
2h 8m 3s q10:export-D114094367
q10:export-D114094367
2h 8m 3s
Add grid-stride + ROCm cap to fused_8bit_rowwise quantize kernels Build FBGEMM GenAI Aarch64 Linux Wheels #10876: Pull request #6109 opened by q10
2h 15m 46s q10:export-D113353144
q10:export-D113353144
2h 15m 46s
Add grid-stride + ROCm cap to generate_vbe_metadata kernel Build FBGEMM GenAI Aarch64 Linux Wheels #10875: Pull request #6108 opened by q10
2h 18m 13s q10:export-D113351683
q10:export-D113351683
2h 18m 13s
2026-08-04 nightly release (5b0850369cca3015d2b11eb196d552889b2ed532) Build FBGEMM GenAI Aarch64 Linux Wheels #10874: Commit cce1106 pushed by pytorchbot
2h 13m 59s nightly
nightly
2h 13m 59s
plumb per-build wave-size flags from CMake into codegen (#6103) Build FBGEMM GenAI Aarch64 Linux Wheels #10873: Commit 5b08503 pushed by meta-codesync Bot
5h 55m 58s main
main
5h 55m 58s
Add ROCm cap to compact_indices and masked_index (default path) (#6106) Build FBGEMM GenAI Aarch64 Linux Wheels #10872: Commit 204fc30 pushed by meta-codesync Bot
3m 20s main
main
3m 20s
Add ROCm cap to compact_indices and masked_index (default path) Build FBGEMM GenAI Aarch64 Linux Wheels #10871: Pull request #6106 opened by q10
5h 35m 13s q10:export-D113351694
q10:export-D113351694
5h 35m 13s
Migrate -c fbcode.platform010_cuda_version to the cuda constraint m… Build FBGEMM GenAI Aarch64 Linux Wheels #10870: Commit fc28489 pushed by meta-codesync Bot
5h 4m 59s main
main
5h 4m 59s
Create no-FBGEMM_FBCODE stub library to fix no-fbcode streamer test Build FBGEMM GenAI Aarch64 Linux Wheels #10869: Pull request #6105 opened by FriedCosey
3h 4m 12s FriedCosey:export-D114193888
FriedCosey:export-D114193888
3h 4m 12s
Migrate -c fbcode.platform010_cuda_version to the cuda constraint modifier (#6100) Build FBGEMM GenAI Aarch64 Linux Wheels #10868: Pull request #6100 synchronize by wenxin0319
2h 55m 11s wenxin0319:export-D114299816
wenxin0319:export-D114299816
2h 55m 11s
Reland: arch-aware FP8 selection via single-instantiation dispatch Build FBGEMM GenAI Aarch64 Linux Wheels #10867: Pull request #6104 opened by q10
2h 16m 40s q10:export-D114094367
q10:export-D114094367
2h 16m 40s
Add grid-stride + ROCm cap to ssd_generate/update_row_addrs kernels (… Build FBGEMM GenAI Aarch64 Linux Wheels #10866: Commit f6c66ba pushed by meta-codesync Bot
2h 37m 59s main
main
2h 37m 59s
Enable Pyrefly for deeplearning/fbgemm (#6099) Build FBGEMM GenAI Aarch64 Linux Wheels #10865: Commit 5d0a3d9 pushed by meta-codesync Bot
1h 16m 1s main
main
1h 16m 1s
[ROCm] plumb per-build wave-size flags from CMake into codegen Build FBGEMM GenAI Aarch64 Linux Wheels #10864: Pull request #6103 opened by jeffdaily
2h 5m 48s jeffdaily:warpsize-4-wavecmake
jeffdaily:warpsize-4-wavecmake
2h 5m 48s