Announcements, 17th September 2015

Eid Al-Adha Holiday, 20-27th September

Please note that there is NO COVERAGE over Eid, but KSL’s supercomputers will be operating. We encourage you to use the next few days in ensuring functionality of your codes and getting yourselves familiarized with the new Cray programming environment, and scheduler – SLURM. It will be useful if you can revisit the material that was distributed during the recently concluded Cray tutorial (https://www.hpc.kaust.edu.sa/training/2015/cray/material/). Please send us your requests by sending an email to help@hpc.kaust.edu.sa. We will respond to your tickets after the Eid break.

XSEDE HPC Workshop: OpenMP

The registration page for the XSEDE HPC Monthly Workshop Series - October 6 - OpenMP session is up.

The portal registration page can be found here:

https://portal.xsede.org/course-calendar/-/training-user/class/435/session/790/register

If there is enough interest, KSL will investigate the possibility of streaming this course live in the Library computer room on 6th October.

KSL Workshop Towards High Efficiency Computing with Allinea

KAUST Supercomputing Laboratory presents the Allinea Software workshop on HPC profiling and debugging: "Towards High Efficiency Computing with Allinea" on 4th October, starting at 9am.

Workshop topics include:

How to profile an application on Shaheen II, and how to debug and identify bottlenecks with Allinea Forge.

How to optimize and monitor application performance using Allinea Performance Reports. Workshop sessions offer attendees the opportunity to practice using the Allinea tools on Shaheen II with either their own applications or provided examples.

Please register at http://www.hpc.kaust.edu.sa/training/2015/allinea/registrations

When: 4th October 2015
Where: KAUST Auditorium between Building 4 and 5, Level 0
What: https://www.hpc.kaust.edu.sa/training/2015/allinea/
Contacts: Bilel Hadri (bilel.hadri@kaust.edu.sa) and Samar Aseeri (samar.aseeri@kaust.edu.sa).

Tip of the week: Writing Output Files

When you submit a job, by default both standard output and standard error are directed to the same file. You can save the output runs with names containing the job ID, as follows :

With SLURM (on Shaheen II Cray XC40)

#SBATCH --output=output_name-%j.out

#SBATCH --error= error_name-%j.err

where the "%j" is automatically replaced by the job allocation number. For job arrays, it is recommended to use the following convention "name-%A_%a.out", where "%A" is replaced by the job ID and "%a" with the array index.

With Loadleveler ( Neser and Shaheen BG/P)

#!/bin/sh

#@ job_name         = hello-world

#@ output           = $(job_name).$(jobid).out

#@ error            = $(job_name).$(jobid).err

Follow us on Twitter

Follow all the latest news on HPC within the Supercomputing Lab and at KAUST, on Twitter @KAUST_HPC.

Previous Announcements

http://www.hpc.kaust.edu.sa/announcements/

Previous Tips

http://www.hpc.kaust.edu.sa/tip/