Skip to content

Actions: pytorch/FBGEMM

Actions

FBGEMM_GPU-ROCm Benchmark

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
fbgemm_gpu_benchmark_rocm.yml
5,005 workflow runs
5,005 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

avx512 based int8 -> bf16 dequantization FBGEMM_GPU-ROCm Benchmark #5726: Pull request #4912 synchronize by seanx92
In progress seanx92:export-D82507938
seanx92:export-D82507938
In progress
Add Static Dispatch Kernels (#4927) FBGEMM_GPU-ROCm Benchmark #5725: Pull request #4936 synchronize by kqfu
1h 11m 2s kqfu:export-D83258808
kqfu:export-D83258808
1h 11m 2s
Add Inference Feature to Skip Pinned Memory Creation (#1948) FBGEMM_GPU-ROCm Benchmark #5724: Pull request #4924 synchronize by Jason-KChen
1h 7m 5s Jason-KChen:export-D83100663
Jason-KChen:export-D83100663
1h 7m 5s
Add Inference Feature to Skip Pinned Memory Creation (#1948) FBGEMM_GPU-ROCm Benchmark #5723: Pull request #4924 synchronize by Jason-KChen
1h 15m 27s Jason-KChen:export-D83100663
Jason-KChen:export-D83100663
1h 15m 27s
forward performance tuning for MI350 FBGEMM_GPU-ROCm Benchmark #5722: Pull request #4925 synchronize by shbiswas834
Action required ROCm:mi350_dev
ROCm:mi350_dev
Action required
Add Inference Feature to Skip Pinned Memory Creation (#1948) FBGEMM_GPU-ROCm Benchmark #5721: Pull request #4924 synchronize by Jason-KChen
16m 19s Jason-KChen:export-D83100663
Jason-KChen:export-D83100663
16m 19s
Add Inference Feature to Skip Pinned Memory Creation (#1948) FBGEMM_GPU-ROCm Benchmark #5720: Pull request #4924 synchronize by Jason-KChen
14m 31s Jason-KChen:export-D83100663
Jason-KChen:export-D83100663
14m 31s
Add Static Dispatch Kernels (#4927) FBGEMM_GPU-ROCm Benchmark #5719: Pull request #4936 synchronize by kqfu
1h 5m 38s kqfu:export-D83258808
kqfu:export-D83258808
1h 5m 38s
Add Static Dispatch Kernels (#4927) FBGEMM_GPU-ROCm Benchmark #5718: Pull request #4936 opened by kqfu
11m 32s kqfu:export-D83258808
kqfu:export-D83258808
11m 32s
MI350X FP8 triton patch FBGEMM_GPU-ROCm Benchmark #5717: Pull request #4889 synchronize by JChunX
1h 11m 34s JChunX:export-D81180838
JChunX:export-D81180838
1h 11m 34s
MI350X FP8 triton patch FBGEMM_GPU-ROCm Benchmark #5716: Pull request #4889 synchronize by JChunX
54s JChunX:export-D81180838
JChunX:export-D81180838
54s
Add Static Dispatch Kernels FBGEMM_GPU-ROCm Benchmark #5715: Pull request #4927 synchronize by kqfu
50s kqfu:export-D83028503
kqfu:export-D83028503
50s
Add Static Dispatch Kernels FBGEMM_GPU-ROCm Benchmark #5714: Pull request #4927 synchronize by kqfu
49s kqfu:export-D83028503
kqfu:export-D83028503
49s
Add head_dim = 64 in B200 Attention. FBGEMM_GPU-ROCm Benchmark #5713: Pull request #4935 opened by gchalump
56m 58s gchalump:export-D82996471
gchalump:export-D82996471
56m 58s
Support bf16 in blackwell cutlass decode attention kernel FBGEMM_GPU-ROCm Benchmark #5712: Pull request #4916 synchronize by Aya-ZIbra
1h 9m 15s Aya-ZIbra:export-D82991495
Aya-ZIbra:export-D82991495
1h 9m 15s
Add Inference Feature to Skip Pinned Memory Creation (#1948) FBGEMM_GPU-ROCm Benchmark #5711: Pull request #4924 synchronize by Jason-KChen
1h 12m 8s Jason-KChen:export-D83100663
Jason-KChen:export-D83100663
1h 12m 8s
Add Inference Feature to Skip Pinned Memory Creation (#1948) FBGEMM_GPU-ROCm Benchmark #5710: Pull request #4924 synchronize by Jason-KChen
12m 37s Jason-KChen:export-D83100663
Jason-KChen:export-D83100663
12m 37s
Split grouped gemm metadata kernel into grouped_common.cuh FBGEMM_GPU-ROCm Benchmark #5709: Pull request #4932 synchronize by cthi
1h 7m 13s cthi:export-D83151150
cthi:export-D83151150
1h 7m 13s
avx512 based int8 -> bf16 dequantization FBGEMM_GPU-ROCm Benchmark #5708: Pull request #4912 synchronize by seanx92
1h 18m 55s seanx92:export-D82507938
seanx92:export-D82507938
1h 18m 55s
FP4 grouped refactor FBGEMM_GPU-ROCm Benchmark #5707: Pull request #4934 opened by cthi
1h 8m 34s cthi:export-D83166227
cthi:export-D83166227
1h 8m 34s
Add output as an option in CUTLASS grouped GEMM FBGEMM_GPU-ROCm Benchmark #5706: Pull request #4931 synchronize by jiawenliu64
57m 55s jiawenliu64:export-D83126291
jiawenliu64:export-D83126291
57m 55s
Fix typo in FP4 quantize FBGEMM_GPU-ROCm Benchmark #5705: Pull request #4933 opened by cthi
58m 40s cthi:export-D83083612
cthi:export-D83083612
58m 40s
Split grouped gemm metadata kernel into grouped_common.cuh FBGEMM_GPU-ROCm Benchmark #5704: Pull request #4932 opened by cthi
1h 8m 9s cthi:export-D83151150
cthi:export-D83151150
1h 8m 9s
more hipify v2 fixes (#4854) FBGEMM_GPU-ROCm Benchmark #5703: Pull request #4921 synchronize by q10
56m 48s q10:export-D82186865
q10:export-D82186865
56m 48s
Migrate GenAI gqa attn splitk kernel arguments to use PTA_B FBGEMM_GPU-ROCm Benchmark #5702: Pull request #4922 synchronize by q10
1h 32m 32s q10:export-D81829444
q10:export-D81829444
1h 32m 32s