[Gate-users] Still working on Gate 9.0

Thu Apr 9 14:58:06 CEST 2020

Hi David and John,

Sorry for the late reply on this thread. I don’t think the collimator implementation is a bug or issue, it appears to just be an incomplete feature which the creator (David Strul) abandoned as he “had the feeling reconstruction did not have an urgent need of collimator information.” I’m not too sure what this all means, but I will continue to just use the more-involved method of collimator definition.

As for the CPU usage, it turns out I was a bit overzealous with my cuts so there was a large amount of data being written due tracking of the excessive interactions.

Regards,

Matthew

From: David Sarrut <david.sarrut at creatis.insa-lyon.fr>
Reply-To: "David.Sarrut at creatis.insa-lyon.fr" <david.sarrut at creatis.insa-lyon.fr>
Date: Thursday, March 12, 2020 at 07:02
To: John Apostolakis <John.Apostolakis at cern.ch>, gate-users <gate-users at lists.opengatecollaboration.org>
Cc: Matthew Strugari <matthew.strugari at dal.ca>
Subject: Re: [Gate-users] Still working on Gate 9.0

CAUTION: The Sender of this email is not from within Dalhousie.
Hi Matthew,

- regarding the first question, could you please fill an issue and attach some macros files to illustrate the problem (I do not personally know it), so that someone can investigate?

- regarding the second point, as John very clearly said, it is unlikely to have a simulation with less than 95% CPU usage (unless of course other if other simulation or process are running in parallel on the same core ;). It may be related to input/output ... is the simulation writing a lot of data?

thanks,
David

On Thu, Mar 12, 2020 at 10:51 AM John Apostolakis <John.Apostolakis at cern.ch<mailto:John.Apostolakis at cern.ch>> wrote:
Dear Matthew,

Do you have an input file which can demonstrate this CPU usage ?  For example do you see it with example SPECT input files ?  If you can specify the GATE version and information about the OS / hardware that you used, it could also be helpful.

In my experience with Geant4 something like this is seen only in applications which have an issue related to input / output, such as saving the Random Number Generator seed too often, or have a problem with multithreading.

It is very unusual for other applications to show less than 95-99% CPU usage.

Best regards,
John
===================================================
John Apostolakis,  EP Department, CERN
SFT (SoFTware for Experiments) Group
Office: Building 32/ room R-003 (ground floor),  Mail: J27210
Email:  john.apostolakis at cern.ch<mailto:john.apostolakis at cern.ch>       Office Tel:  +41-22-767-7239
------------------------------------------------------------------------------------------

On 11 Mar 2020, at 19:20, Matthew Strugari <matthew.strugari at dal.ca<mailto:matthew.strugari at dal.ca>> wrote:

Hi David and Gate-collaborators,

Thank you all for your efforts on the Gate project!

I have a couple questions in regards to the upcoming release. When modelling the collimator, the docs indicate that there is a parameterized collimator setup for parallel hole and fan beam collimators. When executing my macro, I would receive the message "GateVVolume.cc<http://gatevvolume.cc/> (l.192): The material of the volume side hole 1 is not defined." After searching around in the internet for a bit, I came across Dr. Pietrzyk's discussion from 2015 (see here http://lists.opengatecollaboration.org/pipermail/gate-users/2015-July/008546.html) highlighting problems with the SPECThead model/class which does not include a component for the collimator. Has this issue been fixed or will this still be the case with Gate 9.0?

The other general question that I have is: What percentage of CPU usage could I expect with general simulations? I am running simulations in batch mode and when I was performing dose point kernel calculations with the dose actor, the CPU usage was 100% for all 12 physical cores. However, when running SPECT simulations, the CPU usage fluctuates greatly between 10%-75%. I understand that 100% CPU usage is not ideal, but is there anyway to increase CPU usage for SPECT simulations (i.e. overclocking) to speed up the overall simulation duration?

Positively,
Matthew

________________________________
From: Gate-users <gate-users-bounces at lists.opengatecollaboration.org<mailto:gate-users-bounces at lists.opengatecollaboration.org>> on behalf of David Sarrut <David.Sarrut at creatis.insa-lyon.fr<mailto:David.Sarrut at creatis.insa-lyon.fr>>
Sent: March 10, 2020 6:51 AM
To: gate-users <gate-users at lists.opengatecollaboration.org<mailto:gate-users at lists.opengatecollaboration.org>>
Subject: [Gate-users] Still working on Gate 9.0

As you know, we are currently preparing the next Gate release, version 9.0, based on the current Geant4 release 10.6 (patch01).

We will perform the remaining tests in the next few days and plan to tag the 9.0 on early next week. The associated vGate, the virtual machine, will be provided a few days after (thanks Maxime!).
In the meantime, please contact us for any remarks regarding this release, and keep continue to test it with your own favourite simulation.

Also, keep in mind that the development is continuous: any PR that will arrive after this release will be integrated into the develop branch, will remain accessible to everyone and will officially be part of the next release.

Sincerely,
David
PS1: I am sorry, I cannot personally reply to all the emails/questions I receive (and I am not expert in all parts of Gate). I try my best, but, please don't take it badly if I don't reply ...

PS2: I know it is debatable, but it was decided to keep the 9.0 tag instead of 8.3 because this version contains a lot of changes (the policy in the past was not to upgrade the number according to backward compatibility).

--
David Sarrut, Phd
Directeur de recherche CNRS
CREATIS, UMR CNRS 5220, Inserm U1206
Centre de lutte contre le cancer Léon Bérard
28 rue Laënnec, 69373 Lyon cedex 08
Tel : 04 78 78 51 51 / 06 74 72 05 42
http://www.creatis.insa-lyon.fr/~dsarrut
_________________________________
 "2 + 2 = 5,  for extremely large values of 2"
_________________________________
_______________________________________________
Gate-users mailing list
Gate-users at lists.opengatecollaboration.org<mailto:Gate-users at lists.opengatecollaboration.org>
http://lists.opengatecollaboration.org/mailman/listinfo/gate-users

--
David Sarrut, Phd
Directeur de recherche CNRS
CREATIS, UMR CNRS 5220, Inserm U1206
Centre de lutte contre le cancer Léon Bérard
28 rue Laënnec, 69373 Lyon cedex 08
Tel : 04 78 78 51 51 / 06 74 72 05 42
http://www.creatis.insa-lyon.fr/~dsarrut
_________________________________
 "2 + 2 = 5,  for extremely large values of 2"
_________________________________
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opengatecollaboration.org/pipermail/gate-users/attachments/20200409/f2c73706/attachment-0001.html>