For HPC folks, perf events like FLOPs are very interesting. We need to explore how we can monitor them and most importantly how we can maintain them. We have already curated a good part of these events in SDP Perfmon repo. We can start from there as base and try to add new processor families