error in code compilation
lcq... at gmail.com
lcq... at gmail.com
Fri Mar 3 20:04:44 UTC 2017
hi,
i am trying to complete the compilation of the cp2k code version 3.0.0.
Unfortunately, I realized that something went wrong only after I execute
the ./cp2k.popt code. Let me explain.
Throughout the compilation of the cp2k.popt binary, I only observe some
warning messages. In fact, the execution of the "make" command does
generate a cp2k.popt binary. However, when executing the code, I only
observe relevant results whenever I execute the cp2k.popt binary with one
core; executing the same input file with 2 or more cores produces SCF
energies different than the ones expected for this simple system. Note that
I have data from previous runs. To make things more clear, I am attaching
the outcome of the two runs, the one-core execution and the two-core
execution. Some minor differences are ought to be expected; however, as
observed in the attached output files, the SCF energies change dramatically
between the one-core and two-core execution.
Definitely, something is wrong with my compilation; but given the fact that
the compilation runs to completion makes me wonder about what I did wrong,
or is it the case that if the "mpi" variables are not properly defined, a
binary is always generated and this binary cannot run in parallel, for
example?
Any pointers would be appreciated.
Best,
S. Sanchez
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20170303/3e3adeac/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: one_core.output
Type: application/octet-stream
Size: 11402 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20170303/3e3adeac/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: two_cores.output
Type: application/octet-stream
Size: 11807 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20170303/3e3adeac/attachment-0001.obj>
More information about the CP2K-user
mailing list