CI

perf(ggml-cuda): skip shared reduction and barrier for single-wave mu… #1510

Sign in to view logs

Triggered via push July 16, 2026 01:32

davide221

main

Status Success

Total duration 19m 45s

Artifacts –

ci.yml

on: push

uv workspace (lock + sync + import smoke)

Build (cmake + uv sync --extra megakernel)

GPU tests (self-hosted RTX 3090, sm_86)

GPU tests (self-hosted Radeon 8060S, gfx1151 / ROCm)