[slurm-users] seff Not Caluculating [FIXED?]

Jason Simms simmsj at lafayette.edu
Wed Nov 18 14:15:59 UTC 2020


Dear Diego,

A while back, I attempted to make some edits locally to see whether I could
produce "better" results. Here is a comparison of the output of your latest
version, and then mine:

[root at hpc bin]# seff 24567
Use of uninitialized value $hash{"2"} in division (/) at /bin/seff line
108, <DATA> line 602.
Use of uninitialized value $hash{"2"} in division (/) at /bin/seff line
108, <DATA> line 602.
Job ID: 24567
Cluster: hpc
User/Group: chuat/users
State: COMPLETED (exit code 0)
Nodes: 1
Cores per node: 4
CPU Utilized: 00:00:00
CPU Efficiency: 0.00% of 00:00:40 core-walltime
Memory Utilized: 1.24 MB (estimated maximum)
Memory Efficiency: 0.06% of 2.00 GB (2.00 GB/node)

[root at hpc bin]# seff.bak 24567
Job ID: 24567
Cluster: hpc
User/Group: chuat/users
State: COMPLETED (exit code 0)
Nodes: 1
Cores per node: 4
CPU Utilized: 00:00:20
CPU Efficiency: 50.00% of 00:00:40 core-walltime
Job Wall-clock time: 00:00:10
Memory Utilized: 1.24 MB
Memory Efficiency: 0.06% of 2.00 GB

Yours doesn't seem to report anything for CPU Utilized or CPU Efficiency.
At the same time, however, the changes I made to my code to produce those
results may not even be "correct." Moreover, my version may or may not work
for multi-node jobs; I have no way to test those, since at the moment no
user is running them.

For what it's worth, here is a diff of your script vs. mine, in case that's
helpful. That said, while I used to code Perl scripts all the time, I
haven't in, oh, about 20 years, so again, my edits could be entirely the
wrong approach:

https://docs.google.com/document/d/1esg5dE6bP2EzJsLAqIcgfBdgo2GPLKtLNiKoYrdjUl0/edit?usp=sharing

Warmest regards,
Jason

On Tue, Nov 17, 2020 at 6:42 AM Diego Zuccato <diego.zuccato at unibo.it>
wrote:

> Il 09/11/20 12:53, Diego Zuccato ha scritto:
>
> > Seems my corrections actually work only for single-node jobs.
> > In case of multi-node jobs, it only considers the memory used on one
> > node, hence understimates the real efficiency.
> > Someone more knowledgeable than me can spot the error?Seems I managed to
> have it account for the memory on all the nodes.
> See attached file.
> The results seem quite meaningful and match the ones done by hand.
>
> --
> Diego Zuccato
> DIFA - Dip. di Fisica e Astronomia
> Servizi Informatici
> Alma Mater Studiorum - Università di Bologna
> V.le Berti-Pichat 6/2 - 40127 Bologna - Italy
> tel.: +39 051 20 95786
>


-- 
*Jason L. Simms, Ph.D., M.P.H.*
Manager of Research and High-Performance Computing
XSEDE Campus Champion
Lafayette College
Information Technology Services
710 Sullivan Rd | Easton, PA 18042
Office: 112 Skillman Library
p: (610) 330-5632
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.schedmd.com/pipermail/slurm-users/attachments/20201118/6b7c59c5/attachment.htm>


More information about the slurm-users mailing list