[Gate-users] Parallel Processing Limits

Mijat Paunovic paunovicjata at gmail.com
Mon Apr 25 21:39:04 CEST 2022


Dear Gate Community,
I have been using HTCondor on my server with 32 CPUs in order to increase
to computation speed of a simple CT scanner.  I want to run this simulation
for 10^13 X-ray photons per gantry angle. After running incrementally for
10^6,10^7,10^8... I realized that 10^13 would take unreasonable amount of
time since already at 10^10 split on 32 CPUs the computing time is around
15 hours.
Is there a way (besides the variance reduction techniques) to increase the
computing time except simply dividing the total amount of histories by the
number of CPUs? Does gjs do anything besides this?
Kind regards,
Mijat Paunović
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opengatecollaboration.org/pipermail/gate-users/attachments/20220425/2d9a075c/attachment.html>


More information about the Gate-users mailing list