HPCPerfStats

HPCPerfStats copied to clipboard

XALT Data Integration

2

XAlt has yet to be integrated into our new TACC Stats version. This is critical for feature parity with the older versions of TACC Stats.

stephenlienharrell

This will allow us to find failures in a much nicer way, and hide some failures from the user. Also this is the way every major web app has been...

stephenlienharrell

config file processing needs to be dynamic and not only done in setup.py on installation

1

stephenlienharrell

configparser corrections

1

sanga1999

sanga1999

Update monitor to read propper MSRs and memory counters on Ice Lake/Saphire Rapids

2

stephenlienharrell

How many GPUs used counter

2

We need a counter on the new version that says how many GPUs were used for a job.

stephenlienharrell

https://github.com/TACC/tacc_stats/blob/b6da9f3476396670e0d13bf637b5b7c047eaf6fd/monitor/src/main.c#L162C1-L162C1

stephenlienharrell

https://github.com/TACC/tacc_stats/blob/b6da9f3476396670e0d13bf637b5b7c047eaf6fd/monitor/src/main.c#L92C53-L92C55

stephenlienharrell

Array jobs not apprearing

1

From user: We found that when including a slurm array processing line in a slurm batch script, that this prevents any stats information from being recorded. #SBATCH --array=0-1 We created...

stephenlienharrell