Segmentation faults with complex valued kpoints

[Joshua Schiller]

Hello,

I am getting segmentation faults when evaluating complex valued k-points:

cutoff 12.6562

 total centers 229

Ewald sum using 785 reciprocal points

_pmiu_daemon(SIGCHLD): [NID 22108] [c9-5c1s1n2] [Wed Mar 30 11:22:25 2016] PE RANK 323 exit signal Segmentation fault

[NID 22108] 2016-03-30 11:22:25 Apid 36711284: initiated application termination

Application 36711284 exit codes: 139

Application 36711284 exit signals: Killed

Application 36711284 resources: utime ~20s, stime ~468s, Rss ~27028, inblocks ~497027, outblocks ~1846192

als: Killed

Application 36711279 resources: utime ~24s, stime ~470s, Rss ~30212, inblocks ~497969, outblocks ~1846202

2:24 2016] PE RANK 275 exit signal Segmentation fault

_pmiu_daemon(SIGCHLD): [NID 07601] [c12-9c1s7n1] [Wed Mar 30 11:22:24 2016] PE RANK 58 exit signal Segmentation fault

_pmiu_daemon(SIGCHLD): [NID 07606] [c12-9c1s4n0] [Wed Mar 30 11:22:24 2016] PE RANK 222 exit signal Segmentation fault

_pmiu_daemon(SIGCHLD): [NID 07656] [c12-9c1s4n2] [Wed Mar 30 11:22:24 2016] PE RANK 478 exit signal Segmentation fault

_pmiu_daemon(SIGCHLD): [NID 07605] [c12-9c1s5n1] [Wed Mar 30 11:22:24 2016] PE RANK 190 exit signal Segmentation fault

_pmiu_daemon(SIGCHLD): [NID 07659] [c12-9c1s5n3] [Wed Mar 30 11:22:24 2016] PE RANK 446 exit signal Segmentation fault

Application 36711282 exit codes: 139

Application 36711282 exit signals: Killed

Application 36711282 resources: utime ~52s, stime ~470s, Rss ~26984, inblocks ~498326, outblocks ~1846193

Is this a resource issue or some other technical issue? If it is a resource issue, is there a way of estimating memory requirements with qwalk?

Best,

Josh

[Lucas Wagner]

Hi Josh,

As usual, I cannot tell what the problem is without more information than that. If you have a problem with the code, please post a minimal example that exhibits the problem, and some description of what you tested.

For example:

- Did you test on multiple computers?

- Did you test with/without BLAS?

- etc..

Lucas

--
You received this message because you are subscribed to the Google Groups "QWalk user's group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to qwalk-users...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Joshua Schiller]

I didn't test it on multiple computers or without BLAS. I would have thought if it worked on real valued k-points, it should work on complex valued ones. Obviously it does not, but I would not know enough about the code to guess otherwise. Input files are attached.