Skip to content

Actions: pytorch/FBGEMM

Actions

FBGEMM_GPU-CPU Benchmark

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
fbgemm_gpu_benchmark_cpu.yml
5,045 workflow runs
5,045 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

Add ability to pad the rowwise quantized tensors FBGEMM_GPU-CPU Benchmark #5599: Pull request #4877 synchronize by yyetim
20m 13s yyetim:export-D82486197
yyetim:export-D82486197
20m 13s
Migrate GenAI kv cache kernel arguments to use PTA_B FBGEMM_GPU-CPU Benchmark #5598: Pull request #4899 opened by q10
18m 19s q10:export-D82775043
q10:export-D82775043
18m 19s
Restrict to fp8, and patch 4.2.0 release changes FBGEMM_GPU-CPU Benchmark #5597: Pull request #4898 opened by henrylhtsang
18m 48s henrylhtsang:export-D82792657
henrylhtsang:export-D82792657
18m 48s
Add softmax_scale to blackwell cutlass fmha FBGEMM_GPU-CPU Benchmark #5596: Pull request #4897 opened by henrylhtsang
20m 0s henrylhtsang:export-D82788784
henrylhtsang:export-D82788784
20m 0s
Small changes to improve blackwell_fmha_test.py FBGEMM_GPU-CPU Benchmark #5595: Pull request #4896 synchronize by henrylhtsang
19m 59s henrylhtsang:export-D81992869
henrylhtsang:export-D81992869
19m 59s
Small changes to improve blackwell_fmha_test.py FBGEMM_GPU-CPU Benchmark #5594: Pull request #4896 synchronize by henrylhtsang
18s henrylhtsang:export-D81992869
henrylhtsang:export-D81992869
18s
Small changes to improve blackwell_fmha_test.py FBGEMM_GPU-CPU Benchmark #5593: Pull request #4896 opened by henrylhtsang
17m 27s henrylhtsang:export-D81992869
henrylhtsang:export-D81992869
17m 27s
Migrate GenAI kv cache kernels to FBGEMM_LAUNCH_KERNEL, pt 4 FBGEMM_GPU-CPU Benchmark #5592: Pull request #4895 opened by q10
20m 16s q10:export-D82773258
q10:export-D82773258
20m 16s
Migrate GenAI gqa attn splitk kernels to FBGEMM_LAUNCH_KERNEL, pt 1 FBGEMM_GPU-CPU Benchmark #5591: Pull request #4894 opened by q10
21m 30s q10:export-D81817544
q10:export-D81817544
21m 30s
upgrade cutlass to 4.2.0 for fbcode FBGEMM_GPU-CPU Benchmark #5590: Pull request #4893 opened by henrylhtsang
20m 57s henrylhtsang:export-D82567751
henrylhtsang:export-D82567751
20m 57s
Back out "Enable CUTLASS grouped GEMM for pretraining wgrad on GB200 and H100" FBGEMM_GPU-CPU Benchmark #5589: Pull request #4892 synchronize by jiawenliu64
19m 34s jiawenliu64:export-D82746070
jiawenliu64:export-D82746070
19m 34s
Back out "Enable CUTLASS grouped GEMM for pretraining wgrad on GB200 and H100" FBGEMM_GPU-CPU Benchmark #5588: Pull request #4892 opened by jiawenliu64
10m 59s jiawenliu64:export-D82746070
jiawenliu64:export-D82746070
10m 59s
add blog link FBGEMM_GPU-CPU Benchmark #5587: Pull request #4888 synchronize by sijiac
20m 22s sijiac:export-D82681043
sijiac:export-D82681043
20m 22s
Add ability to pad the rowwise quantized tensors FBGEMM_GPU-CPU Benchmark #5586: Pull request #4877 synchronize by yyetim
19m 31s yyetim:export-D82486197
yyetim:export-D82486197
19m 31s
dequantize_fp8_cache_kernel: Move D=128 device-side-assertion check to host FBGEMM_GPU-CPU Benchmark #5585: Pull request #4869 synchronize by ColinPeppler
21m 42s ColinPeppler:export-D82320518
ColinPeppler:export-D82320518
21m 42s
dequantize_fp8_cache_kernel: Move D=128 device-side-assertion check to host FBGEMM_GPU-CPU Benchmark #5584: Pull request #4869 synchronize by ColinPeppler
19s ColinPeppler:export-D82320518
ColinPeppler:export-D82320518
19s
symmetric quantization to FBGEMM prefill token-wise FP8 (fixed) FBGEMM_GPU-CPU Benchmark #5583: Pull request #4868 synchronize by ColinPeppler
18m 13s ColinPeppler:export-D82320500
ColinPeppler:export-D82320500
18m 13s
fix the scaled input issue FBGEMM_GPU-CPU Benchmark #5582: Pull request #4884 synchronize by pyjhzwh
19m 4s pyjhzwh:export-D82147819
pyjhzwh:export-D82147819
19m 4s
Add ability to pad the rowwise quantized tensors FBGEMM_GPU-CPU Benchmark #5581: Pull request #4877 synchronize by yyetim
19m 51s yyetim:export-D82486197
yyetim:export-D82486197
19m 51s
Add ability to pad the rowwise quantized tensors FBGEMM_GPU-CPU Benchmark #5580: Pull request #4877 synchronize by yyetim
6m 49s yyetim:export-D82486197
yyetim:export-D82486197
6m 49s
Backward optimizations + rocm and gfx950 build support FBGEMM_GPU-CPU Benchmark #5579: Pull request #4729 synchronize by liligwu
Action required ROCm:shbiswas/uptream_develop_0815
ROCm:shbiswas/uptream_develop_0815
Action required
Optimize wgrad CUTLASS grouped gemm FBGEMM_GPU-CPU Benchmark #5578: Pull request #4891 opened by jiawenliu64
19m 23s export-D82700455
export-D82700455
19m 23s
Support multiple total-k and total-m in quantize bench FBGEMM_GPU-CPU Benchmark #5577: Pull request #4890 opened by jiawenliu64
18m 2s jiawenliu64:export-D82700396
jiawenliu64:export-D82700396
18m 2s
Add ability to pad the rowwise quantized tensors FBGEMM_GPU-CPU Benchmark #5576: Pull request #4877 synchronize by yyetim
19m 55s yyetim:export-D82486197
yyetim:export-D82486197
19m 55s
fix the scaled input issue FBGEMM_GPU-CPU Benchmark #5575: Pull request #4884 synchronize by pyjhzwh
19m 56s pyjhzwh:export-D82147819
pyjhzwh:export-D82147819
19m 56s