Unusual memory consumption

Dear Allpix Squared team,

I am trying to simulate a laser beam deposition using the modules DepositionLaser and TransientPropagation (see the configuration file attached). The geometry is just one simple detector comparable to a diode. When I try to run this on my own machine, the memory use goes up to 8GB and my program is killed (see photo). I think it might have something to do with the output plots of TransientPropagation. When I turn off the output plots, it runs just fine and no more than 500MB of memory is used. What could be the cause of this?

Perhaps relevant, when I ran it on lxplus, I got the error “Error in TBufferFile::WriteByteCount: bytecount too large (more than 1073741822)”. Often it was then killed, but a few times it was able to finish running (with this ByteCount error) and I was able to open it. I do not understand why it worked a couple of times. When I was able to open the ROOT file, I could not open the induced_charge_e_vs_depth_histo, induced_charge_h_vs_depth_histo or the induced_charge_map, because ROOT would crash. This is also a reason why I think it has to do with the output_plots in TransientPropagation.

Lastly, I am pretty sure I am using the most recent version of allpix squared, which I cloned from gitlab today, but when I do allpix --version, I see v2.0.0 (see photo). I don’t know if it could be relevant but I thought I should mention it.

Thanks in advance!

Femke

allpix_version

configuration_file_memory_error.txt (716 Bytes)

Update:
With some help and the use of a debugger we have found out that the error was that the pixel size used was too large (1cm 1cm), which created histograms with a very large amount of bins.

Problem solved :slight_smile:

Hi @fdewit

ha, you were quicker than we could even answer :slight_smile: Nice that you found the issue - I would gave pointed in a similar direction. We generate some pretty detailed plots there, and with many bins the memory allocated by ROOT for these just explodes…

Best,
Simon

Yes, exactly. Especially the 2D plots took a lot of memory. Thank you and sorry for taking your time on this!