Triton benchmarks · Workflow runs · intel/intel-xpu-backend-for-triton · GitHub

Actions

All workflows
Workflows
- Build and test Build and test
- E2E accuracy E2E accuracy
- PyTorch inductor tests on Linux PyTorch inductor tests on Linux
- Triton benchmarks Triton benchmarks
- Triton wheels Triton wheels
- Automate update Translator cid Automate update Translator cid
- Bandit Check Bandit Check
- Build and test gemm Build and test gemm
- Build and test GPU Build and test GPU
- Build and test on Windows Build and test on Windows
Management
- Caches

Triton benchmarks

Actions

Loading...
Loading

triton-benchmarks.yml

107 workflow run results

107 workflow run results

Triton benchmarks Triton benchmarks #717: Scheduled

54m 54s main

main

54m 54s

[XeTLA] Use stream-k implementation by default for 3072x4096x3072 Triton benchmarks #716: Pull request #2496 opened by whitneywhtsang

50m 13s whitneywhtsang/streamk

whitneywhtsang/streamk

50m 13s

[Benchmark] Run xetla streamk gemm in benchmark Triton benchmarks #715: Pull request #2438 synchronize by whitneywhtsang

50m 35s yudong/xetla_streamk

yudong/xetla_streamk

50m 35s

[Benchmark] Run xetla streamk gemm in benchmark Triton benchmarks #714: Pull request #2438 synchronize by etiotto

50m 5s yudong/xetla_streamk

yudong/xetla_streamk

50m 5s

Triton benchmarks Triton benchmarks #713: Manually run by anmyachev

54m 41s amyachev/bench-time

amyachev/bench-time

54m 41s

[Benchmark] Run xetla streamk gemm in benchmark Triton benchmarks #712: Pull request #2438 synchronize by yudongsi

59m 8s yudong/xetla_streamk

yudong/xetla_streamk

59m 8s

Triton benchmarks Triton benchmarks #711: Scheduled

49m 30s main

main

49m 30s