Dear all,
I' m running yambo to solve the Bethe Salpeter eq. on a k points mesh 30x30x1 (nkpt=452), but the code gives the following error message:
"P01: [ERROR] STOP signal received while in :[08.02] Main loop
P01: [ERROR][NetCDF] No such file or directory"
However using a smaller mesh (16x16x1) the code does not give any problems.
Is the problem related to the dimension of the files, or to the disk space?
[ERROR][NetCDF]
Moderators: myrta gruning, andrea marini, Daniele Varsano, Conor Hogan
- myrta gruning
- Posts: 240
- Joined: Tue Mar 17, 2009 11:38 am
- Contact:
Re: [ERROR][NetCDF]
Yes, it may be a problem related to the fact the ndb.* becomes larger than 2GB. I also got strange messages from the Netcdf in that case (even though this seems new to me). I am not a netcdf guru, but as I said I experienced some similar chrashes, so here some suggestions, I hope they may help you.cudazzo wrote:Dear all,
I' m running yambo to solve the Bethe Salpeter eq. on a k points mesh 30x30x1 (nkpt=452), but the code gives the following error message:
"P01: [ERROR] STOP signal received while in :[08.02] Main loop
P01: [ERROR][NetCDF] No such file or directory"
However using a smaller mesh (16x16x1) the code does not give any problems.
Is the problem related to the dimension of the files, or to the disk space?
One way to prevent problems with large ndb.* (usually) is to invoke Yambo with the -S keyword that fragmented the database (yambo -H).
However, if you did not use this keyword from the beginning and you have a RESTART directory (and the calculations took already some time so that starting from scratch is not a good option) restarting with -S may cause a little mess (at least I was able to generate a mess in such a case).
Another option (maybe safer in the restart case) is to reconfigure Yambo with --enable-largedb feature (for more info ./configure --help in the yambo dir) that should enable netcdf to work with files larger than 2GB (even though I do not know exactly if this is possible only for 64 bits archs, and how this may affect performance, I/O etc.), and see if the program get over the point where it crashes now.
Cheers,
m
Dr Myrta Grüning
School of Mathematics and Physics
Queen's University Belfast - Northern Ireland
http://www.researcherid.com/rid/B-1515-2009
School of Mathematics and Physics
Queen's University Belfast - Northern Ireland
http://www.researcherid.com/rid/B-1515-2009
- Conor Hogan
- Posts: 111
- Joined: Tue Mar 17, 2009 12:17 pm
- Contact:
Re: [ERROR][NetCDF]
Another common reason for this error is that the netCDF databases have not been created properly for some reason. Check that the SAVE directory contains a directory wf_0000n_... for every k-point used, and particularly that the size of all the SAVE/wf*/ndb.fragment files are about the same. It may simply be that you have to run a2y/p2y again.
Dr. Conor Hogan
CNR-ISM, via Fosso del Cavaliere, 00133 Roma, Italy;
Department of Physics and European Theoretical Spectroscopy Facility (ETSF),
University of Rome "Tor Vergata".
CNR-ISM, via Fosso del Cavaliere, 00133 Roma, Italy;
Department of Physics and European Theoretical Spectroscopy Facility (ETSF),
University of Rome "Tor Vergata".