0
Fork 0
mirror of https://github.com/zml/zml.git synced 2026-11-16 19:11:09 +00:00
zml/bin/zml-smi/info
Alhussein Jamil ecb824e859
bin/zml-smi: keep NVML polling on one thread so --top does not freeze (#766)
## Summary
- `zml-smi --top` was stalling under load because each GPU and the
process list hit NVML on their own thread. Those calls serialize on the
driver RM lock, so one slow ioctl froze the whole TUI.
- All NVIDIA queries now run on a single poller.
`nvmlDeviceGetPcieThroughput` (it blocks while sampling) is only taken
on the first tick and every fourth tick after that; link speed/width
stay on the hot path.
- Process-query sizes above 4096 are ignored, and a failed per-process
util sample no longer skips the rest of that GPU.

---------

Co-authored-by: alhussein jamil <alhussein@zml.ai>
History 2026-09-11 16:29:29 +02:00
..
BUILD.bazel bin/zml-smi: Add CSV export and driver version reporting (#452) 2026-04-07 16:44:52 +02:00
device_info.zig bin/smi: add oneapi smi monitoring (#557) 2026-05-29 18:14:38 +02:00
host_info.zig bin/zml-smi: Add CSV export and driver version reporting (#452) 2026-04-07 16:44:52 +02:00
info.zig bin/zml-smi: Add CSV export and driver version reporting (#452) 2026-04-07 16:44:52 +02:00
poll_metrics.zig bin/zml-smi: keep NVML polling on one thread so --top does not freeze (#766) 2026-09-11 16:29:29 +02:00
process_info.zig bin/zml-smi: add prometheus endpoint, json and macos host infos (#455) 2026-04-10 17:26:09 +02:00