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!



configuration_file_memory_error.txt (716 Bytes)

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…


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