Page 1 of 1

SIGSEGVProgram - another case

Posted: 01 Feb 2012, 19:22
by jtobin
Hello,
I am getting Terminated by SIGSEGVProgram trace-back ....when i run the elegant code.
The code uses a script - "Multislits.tcl" to simulate a transverse slit along the beam. When I run the code without this script, things are fine.
And the script itself is fine when run with an input file. But when i call it using the lte file from ele file , elegant gives this error.
I tried to read through other SIGSEGV issues people have encountered before in the forum. Most of them were associated with a BIN issue or a missing parameter.
Any idea why this might be happening?
regards,
Charles
(Btw, I am unable to post these files to the forum…it keep giving me an error stating Could not upload attachment to ./files/84_534236bafa947e2232c48add8e7a3d56.)

Re: SIGSEGVProgram - another case

Posted: 02 Feb 2012, 14:10
by michael_borland
Charles,

There is a bug in elegant related to particle accounting after a SCRIPT element in which particles are lost. You can work around this by setting use_particle_id=0 and drift_matrix=1 on your script element.

We'll look into the file upload issue.

--Michael

Re: SIGSEGVProgram - another case

Posted: 02 Feb 2012, 14:19
by jtobin
Thank you Michael. It works now! Thanks...

regards,
Charles

Re: SIGSEGVProgram - another case

Posted: 02 Feb 2012, 15:15
by michael_borland
Charles,

Another workaround, due to Yusong Wang, is to ask for lost particle output by giving losses=%s.los in &run_setup. (The bug is that the code assumes you always ask for this.) That way, you can remove the use_particle_id=0 parameter on the SCRIPT element and the particle IDs will be consistent throughout the output files.

--Michael