[CP2K:4790] Re: Trying to compile cp2k+cuda+intel_mpi in parallel-smp mode

Urban Borštnik urban.b... at gmail.com
Mon Nov 18 10:06:23 CET 2013


Hi Alexey,

One trick you could try is:
        mpirun [mpi options] gdb --batch --eval-command=run --eval-command=bt --nw --args /path/to/cp2k.popt -i input_file -o output_file"
You might have to modify the "run" and "bt" command sequence to catch
the right behavior when you use threads (i.e., to print the backtrace
from all the threads).

You could also do something similar with valgrind's thread correctness
checker.

Best regards,
Urban
        
On Mon, 2013-11-18 at 00:06 -0800, Alexey Nesternko wrote:
> I'm trying to solve this problem by myself but I faced some problems
> via debugging. It's hard to invade with gdb to the process (which
> makes a double memory free) to determine position in the code. Is
> there anybody here who can give an advice in debugging of hybrid (mpi
> +openmp) programs on the cluster?
> 
> -- 
> You received this message because you are subscribed to the Google
> Groups "cp2k" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to cp2k+uns... at googlegroups.com.
> To post to this group, send email to cp... at googlegroups.com.
> Visit this group at http://groups.google.com/group/cp2k.
> For more options, visit https://groups.google.com/groups/opt_out.





More information about the CP2K-user mailing list