feat: v3.10 GPU stress and NCCL burn updates

This commit is contained in:
Mikhail Chusavitin
2026-03-31 11:22:27 +03:00
parent 6dee8f3509
commit c850b39b01
12 changed files with 423 additions and 88 deletions
+16
View File
@@ -61,6 +61,20 @@ func buildNvidiaStressJob(opts NvidiaStressOptions) (satJob, error) {
collectGPU: true,
gpuIndices: selected,
}, nil
case NvidiaStressLoaderNCCL:
cmd := []string{
"bee-nccl-gpu-stress",
"--seconds", strconv.Itoa(opts.DurationSec),
}
if len(selected) > 0 {
cmd = append(cmd, "--devices", joinIndexList(selected))
}
return satJob{
name: "03-bee-nccl-gpu-stress.log",
cmd: cmd,
collectGPU: true,
gpuIndices: selected,
}, nil
default:
return satJob{}, fmt.Errorf("unknown NVIDIA stress loader %q", opts.Loader)
}
@@ -78,6 +92,8 @@ func normalizeNvidiaStressOptions(opts *NvidiaStressOptions) {
opts.Loader = NvidiaStressLoaderBuiltin
case NvidiaStressLoaderJohn:
opts.Loader = NvidiaStressLoaderJohn
case NvidiaStressLoaderNCCL:
opts.Loader = NvidiaStressLoaderNCCL
default:
opts.Loader = NvidiaStressLoaderBuiltin
}