Dear Slave,
Sorry for my late response, I was inquiring about it in the lab, and was waiting for
feedback for some time. Overall, we are mostly using the GPUs, and there will be an
increase in demand as we and others (Martin Urschler at the Institute of Medical
Informatics) are increasingly moving into machine learning. As such, more GPUs would be
welcome.
Regarding CPUs, we are not using the MedBioNode for CPU computing. Here, either we use our
workstations (<32 cores), or we go to the Vienna Scientific Cluster (>128 cores).
For the segment in between, the extra effort of setting this up on the MedBioNode does not
pay off. Moreover, the cluster is not low latency (infiniband or so), and as such
scalability across nodes is limited.
In short, more GPUs would be helpful - G
Am 01.02.2024 um 15:42 schrieb Trajanoski, Slave <slave.trajanoski(a)medunigraz.at>at>:
Dear MedBioNode user,
it is planned in the next time to expand our HPC infrastructure and therewith to provide
more CPU and GPU performance for jobs execution. To proper plan cluster growth it is very
critical to evaluate your needs, as well as expected amount of future CPU based or GPU
based data analysis tasks.
Therefore provide us information about bottlenecks that you currently experience in using
MedBioNode. Also it will be very helpful if you can already estimate calculation capacity
separately CPU or GPU for your data analysis for this year. If you have some questions, or
you’re uncertain how to provide this information contact me to assist you.
Thanks very much in advance for your cooperation.
Best regards,
Slave
--
Slave Trajanoski, Phd
Senior Scientist Bioinformatics
CF Computational Bioanalytics, Center for medical research
Medical University Graz
Neue Stiftingtalstraße 6-West, Turm/Stiege:P, 4.Stock
8010 Graz
Tel. +43 316 385 73024
E-Mail: slave.trajanoski@medunigraz.at<mailto:slave.trajanoski@medunigraz.at>
_______________________________________________
Medbionode mailing list --
medbionode@lists.medunigraz.at<mailto:medbionode@lists.medunigraz.at>
To unsubscribe send an email to
medbionode-leave@lists.medunigraz.at<mailto:medbionode-leave@lists.medunigraz.at>
--
Gernot Plank, PhD
Professor of Computational Cardiology
Medical University of Graz
Gottfried Schatz Research Center : Biophysics
Neue Stiftingtalstrasse 6/D04
A-8010 Graz
Austria
Phone: +43 316 385 71526
Fax: +43 316 385 79611
Email: gernot.plank(a)medunigraz.at
Web:
http://forschung.medunigraz.at/fodok/staff?name=GernotPlank
http://ccl.medunigraz.at