Learning Forum

[helen.durand]

Hello!

I have a question about one of the outputs from running ANSYS structural on a HPC cluster. The HPC cluster uses ANSYS 2020R1, a Slurm job-scheduler, no security programs are installed, and python 3.7 is being used.

In the log file, it checks UserCPU. Initially for Main it shows 37 and then for server_select is shows 54 UserSPU. That number changes over time at these CHECKINs.

Time:   Main:   server_select

6h         37         54

12h      41         55

18h      41         55

1d         39         58

1d 6h   39         58

1d 12h 37         60

1d 18h 36         62

2d         36         62

2d 6h   35         63

2d 12h 34         64

2d 18h 34         64

3d         33         65

3d 6h   32         66

3d 12h 32         66

3d 18h 32         65

What do these numbers mean and why are they changing?

Thank you!

[mrife]

what log file are you referring to? "Checkin" is a licensing term i.e. check-out of a license and check-in of a license back to the license server. It could also be a Slurm term. If the log is from Slurm, you will need to ask the cluster admin.
Mike

[helen.durand]

Thank you for the reply.
Yes, you are correct that this is dealing with Slurm. This was actually something the server admin was wondering. If you have any ideas on what it means it would be appreciated.
Thanks again!

[mrife]

Please ask this of the cluster admin. MIke

[helen.durand]

Thank you for your help, Mike.

We checked with the Cluster Admin, and here is the information he provided:
That info is not from a Slurm file but rather from an ansys logfile. It is located in:~$USER/.ansys/local.$NODENAME.$UNIQUE_ID.logAre the numbers the total number of licenses that the job has checked out? In other words, could these jobs be hitting licensing limits.
Thank you.