7 comments

[ 3.0 ms ] story [ 30.3 ms ] thread
Renaming fopen64 to intercept library calls feels like a brittle hack masquerading as "sandboxing." Why not just upstream this hardware support to nvtop instead of fragmenting the ecosystem?
If this logic were pushed into nvtop, wouldn't the codebase become unmaintainable? Each vendor's interception method is going to be different.
"NPU" seems to refer to trainium only?
Is it capable of exposing metrics in Prometheus format?
would be nice to have cpu usage added so I have all in one?

currently I use btop which shows basic gpu load along with cpu, network, etc.