Hello,
With A100, on version 3.1.0 the device usage % is correctly shown, but on version greater it does not (> 3.2.0) as it shows N/A.
We discovered that with 3.2.0 and I just tested it again with 3.3.2
The GPU usage is correctly shown on the graph though (drawn at 100%).
This works correctly on H100.
I'll note that all tests passed on a node with A100s.
Any idea on what's needed to fix this?
Happy to provide info/help on resolving this.
Cheers,
Hello,
With
A100, on version3.1.0the device usage % is correctly shown, but on version greater it does not (> 3.2.0) as it showsN/A.We discovered that with
3.2.0and I just tested it again with3.3.2The GPU usage is correctly shown on the graph though (drawn at 100%).
This works correctly on
H100.I'll note that all tests passed on a node with A100s.
Any idea on what's needed to fix this?
Happy to provide info/help on resolving this.
Cheers,