Search results

[amd] MI350 tests are failing

Status: Open.

#1036 In meta-pytorch/tritonbench;

· xuzhao9 opened on Apr 23, 2026

_do_bench_cudagraph_with_cache_clear segfaults on ROCm when DEFAULT_REP=3000ms causes pathological n_repeat

module: rocm

Status: Open.

#1034 In meta-pytorch/tritonbench;

· umechand-amd opened on Apr 23, 2026

[feature request] Support running multiple modes in one process

Status: Open.

#902 In meta-pytorch/tritonbench;

· xuzhao9 opened on Feb 27, 2026

Run "all-or-nothing" mode for avg metrics in the CI

Status: Open.

#753 In meta-pytorch/tritonbench;

· xuzhao9 opened on Dec 18, 2025

[power] Add latency histogram

Status: Open.

#670 In meta-pytorch/tritonbench;

· xuzhao9 opened on Nov 25, 2025

Inspect Tritonparse reproduce errors

Status: Open.

#662 In meta-pytorch/tritonbench;

· xuzhao9 opened on Nov 22, 2025

Add comm and distributed kernels

Status: Open.

#636 In meta-pytorch/tritonbench;

· xuzhao9 opened on Nov 11, 2025

[AMD][AITER] Add more AMD kernels

Status: Open.

#604 In meta-pytorch/tritonbench;

· xuzhao9 opened on Oct 30, 2025

Increase operator coverage

help wantedExtra attention is neededExtra attention is needed

Status: Open.

#479 In meta-pytorch/tritonbench;

· xuzhao9 opened on Sep 26, 2025

profiler + cudagraph hangs forever for rope

Status: Open.

#413 In meta-pytorch/tritonbench;

· BoyuanFeng opened on Sep 10, 2025

[BE] Enable CUDAGraph by default

be

help wantedExtra attention is neededExtra attention is needed

Status: Open.

#349 In meta-pytorch/tritonbench;

· xuzhao9 opened on Aug 21, 2025

Request for branches in repo

Status: Open.

#306 In meta-pytorch/tritonbench;

· bdbowyer opened on Jul 28, 2025