I get this after the Mulliken charges print and Hirshfeld analysis begins:
Performing Hirshfeld population analysis.
Hirshfeld: reading atomic densities from disk.
Q-Chem fatal error occurred in module libmdc/newfileman.C, line 376:
FileMan error: End of file reached prematurely reading (1361584928) bytes in file FILE_HIRSHFELD_BLOCKDM
Path: /work/qcscratch/GGAA3_qc_job_8.5_CHIRS/736.0
If I had to guess, I might suspect that MAX_SCF_CYCLES=0 is causing Q-Chem to skip some necessary setup steps. In any case, it would be useful to find the minimal set up rem variables that causes this crash.
Thanks for the reply. I eliminated the MAX_SCF_CYCLES=0 and reran the job. It failed with a command line output of:
FileMan error: End of file reached prematurely reading (1361584928) bytes in file FILE_HIRSHFELD_BLOCKDM
Path: /work/qcscratch/GGAA3_qc_job_8.5_CHIRS/736.0: No such file or directory
Error in Q-Chem run part 1
Error in the Q-Chem run
The output file from Q-Chem ended with this message:
Performing Hirshfeld population analysis.
Hirshfeld: reading atomic densities from disk.
Q-Chem fatal error occurred in module libmdc/newfileman.C, line 376:
FileMan error: End of file reached prematurely reading (1361584928) bytes in file FILE_HIRSHFELD_BLOCKDM
Path: /work/qcscratch/GGAA3_qc_job_8.5_CHIRS/736.0
Eliminating SCF_GUESS gets us past the error message reported previously. However, after 36 hours (wall) the only Hirshfeld output we get is:
Performing Hirshfeld population analysis.
Hirshfeld: reading atomic densities from disk.
Hirshfeld: performing integration.
warning: vdwdata pointer should have been initialzed before dftman
Nothing else. So, it appears that either the Hirshfeld analysis is being performed but will consume an endless amount of CPU time, or the software is hung.
For a future release, it would be useful for Q-Chem to display progress information so the user can tell if the program is working.
Hi Jim,
I’m not sure what scientific objective you are trying to accomplish but it seems that most of your jobs basically come down to “run Q-Chem on 1000-atom systems and see what breaks”. That might be useful data to have, but please understand that such jobs are not really in the wheelhouse of our customer base, therefore are unlikely to be a priority for our (very small) staff. If the jobs are unrelated to any scientific application then they are unlikely to be a priority for our academic developers either. That said, anyone who has purchased a Q-Chem license can become a developer (i.e., gain access to source code) by signing a non-disclosure agreement. If you are interested in becoming a developer, please contact Q-Chem support (email address on the Q-Chem website, I don’t want to post it here). If you decide to go that route, then either myself of someone at Q-Chem could probably arrange a meeting to discuss what aspects of the code you would like to address.
I am not trying to run and see when Q-Chem breaks. It is unfortunate that you have that impression.
Our work at KressWorks is targeted toward the understanding and cure of Cancer, in particular Ewing Sarcoma which took the life of my sister. Briefly, that work requires exhaustive analysis of simulations of DNA oligomers at the atomistic level, with Molecular Dynamics as well as Quantum Chemistry and Kinetics. The reason our systems are so large is due to the fact that the DNA oligomers we must examine are 40 or so base pairs with explicit H2O, neutralizing ions as well as NaCl solute at 0.2 M concentration.
A more detailed description and discussion (a bit dated but useful) can be found here:
So, I am not attacking Q-Chem. As a matter of fact I am going to have to eat crow and sign up to continue my QMP relationship with Q-Chem. This is the case because Q-Chem has shown, through my continuing use of it, to be a good platform for what I am trying to accomplish.
As I have done in the past with other computational resources that we have used, I would be very interested in securing a developer relationship with the Q-Chem team. Please let me know what I need to do to accomplish that.