|Message: Re: Who's responsible for deletion of hits and hits collection objects?||Not Logged In (login)|
Click on the Forum title, e.g. on the "Forums by Category" page, to read a sequence of postings to the Forum and its threads all in one page. If you are only interested in one thread or the thread following a specific posting, click the thread or the posting, which takes you to a smaller page, which contains only the part you are interested in and may be easier to navigate.
Messages are "chained" if there are only replies at the first level, i.e. 1/1.html, 1/1/1.html etc. In case of "chained" messages the message number is replaced by the icon and there is no indentation.
Inline: Display the subject line only or also the text of the posting(s); for the choice "All" the "Outline" choices are switched off.
|1||0||1||no text / full text of posting|
|2||1||All||text for level 1 only / text for All postings|
Outline: Choose the depth of the posting thread, successive toggle controls provide increasing detail.
|1||2||1||2 levels / 1 level (original posting)|
|2||3||2||3 levels / 2 levels|
|3||3||All||3 levels / all levels (all postings)|
From: "Kazuyoshi Furutaka" <email@example.com> Subject: Re: Who's responsible for deletion of hits and hits collection objects? Date: Fri, 11 Mar 2005 02:56:11 -0800 > > >> Sad thing is, the exampleN0 as well as A01app, when run with > > >> hundreds of millions of events, too are always killed by oom-killer > > >> on my linux system (Fedora Core3, kernel 2.6.10-1.770_FC3, gcc-3.2.4, > > >> libstdc++-3.4.2), and I think something is wrong... > > > > As far as know, exampleN03 does not use SensitiveDetector and > > HitCollections. > > Based on the fact that Michel pointed out and Asai-san's comment that > all hits and HitsCollection objects are surely deleted when G4Event is > deleted at the end of the event, it seems there's another reason for > the out-of-memory. > > To add more information, I've started A01app again on gdb with debug > on (w/ -g), and it's runnning longer than that without '-g' flag > (though it's terribly slow and # of events is still a little more than > 570000 for two days). So it may have something to do with the code > produced by g++. Well... On a Fedora Core 3 system with the latest updates and w/o G4DEBUG, A01app was killed by oom-killer when it processed ~360,000 events. (As I remember it could process as much as 1,000,000 events on a Fedora Core 2 system.) On the same system w/ G4DEBUG=1, A01app could process more than ~920,000 events without being killed. (I made a mistake to say that it was run on gdb; NO, I just executed an A01app w/ G4DEBUG=1 from a command line). On a SLC3 system with the latest updates and w/o G4DEBUG, running as a guest of VMware on the above FC3 system, also, more than ~860,000 events could be processed by A01app. So now I think the OOM is caused by something other than Geant4 (e.g. g++, libstdc++, or glibc)... Is there any way to resolve the OOM issue??? Yours, Kazuyoshi
|Inline Depth:||Outline Depth:||Add message:|