You are here

Relax with RDC data gives segmentation fault

2 posts / 0 new
Last post
Relax with RDC data gives segmentation fault


When I try to use RDC data together with structure refinement (relax), it faults (see the command I tried below). 

relax.linuxgccrelease -in:file:s test.pdb -in:file:rdc test.rdc -score:weights talaris2013 -score:patch rdc_patch -out:pdb -nstruct 20

The rdc_patch has rdc score term. 

The segmentation fault appears when we use talaris2014 or ref2015 score wts as well. I tried this with a different set of  PDB and RDC files, I see the same issue.


Post Situation: 
Mon, 2018-07-23 11:39

Sorry about that -- segmentation faults difficult to track down (and are 100% our fault -- they're most often caused by unexpected input, but at the very least we should provide a decent error message, rather than an obtuse segfault).

My first recommendation would be to compile the debug mode version of Rosetta (add `mode=debug` to the scons commandline, instead of `mode=release`). This adds a bunch of extra checks to the running. This slows Rosetta down by quite a bit, but it often results in an error message instead of the crash. You'll probably not be able to interpret the error message yourself, but you can post it here and we might be able to puzzle it out.

If you still get a segfault with the debug mode, the next step would be to run the program under a debugger and get a backtrace. Alternatively, you might be able to use the `catchsegv` program to get a backtrace (just add `catchsegv` before the command that results in a segfault).

Another option, if you're up to it, is to just post all the relevant input files here (so test.pdb, test.rdc and your rdc_patch file) and I can try debugging it on my machine.

Fri, 2018-08-03 07:54