Dear QE users,
I am using the development version of QE 6-2-2 (downloaded the 05/18)
compiled with Intel 2017.5
I am working with Centos7 on a bi Xeon Skylake.
I have one comment and one question:
I have a problem using PP for making cube file of the charge density
since an update of the OS last week:
If I use use PP only with openmp, the program crash saying it can't
find/open ***.wfc1
it was working (same input, same command) before the update.
I recompile PP, the problem is still there.
I think the problem comes from Intel compiler with the last kernel and
library. I noticed a error message with the vtune amplifier.
( related to
https://software.intel.com/en-us/forums/intel-vtune-amplifier-xe/topic/779038
?)
I don't have time to check it now.
The solution I found, if it can help someone, is simply to run PP with
MPI... It works...
Second, I have a question for dev:
If I make a cube file with pp:
with option 17 (valence charge density) or 21 (with core correction) the
only temporary file is tmp.pp which is normal.
with option 8 (elf, with the patch
https://gitlab.com/QEF/q-e/issues/34) n ***.wfc files (where n is the
number of MPI process used) are created and left in the working directory.
My question is:
Is it normal to have the ***.wfc files with elf? Or could it be a
problem related to the evolution of the code as the problem with
dffts/dfftp corrected by the patch?
Best regards,
Bruno
_______________________________________________
users mailing list
[email protected]
https://lists.quantum-espresso.org/mailman/listinfo/users
