I got a problem with frequency calculation using Gaussian.
- For some molecules, I used opt+freq command --> the opt was done, showing (yes, yes, yes, yes in force and displacement convergence). However, the freq calculation can't be done completely, and the job terminated without any error reported (no convergent failed notice; no momory limited notice; no nothing).
- I set up other freq jobs, using the above optimized structure, then same result; job terminated with out any notice.
I suspected it was due to limited memory, then i clear several files, and sure i have enough space on my hard disk.
Any one got that problem before? please advice!
--------------------------------------------
Ending of the out file:
**** Warning!!: The smallest alpha delta epsilon is 0.25296473D-01
**** Warning!!: The smallest beta delta epsilon is 0.54826547D-01
1 Symmetry operations used in ECPInt.
ECPInt: NShTT= 18528 NPrTT= 119541 LenC2= 15580 LenP2D= 52920.
LDataN: DoStor=T MaxTD1= 8 Len= 415
Number of processors reduced to 5 by ecpmxn.
Symmetrizing basis deriv contribution to polar:
IMax=3 JMax=2 DiffMx= 0.00D+00
G2DrvN: will do 56 centers at a time, making 1 passes doing MaxLOS=2.
Calling FoFCou, ICntrl= 3107 FMM=F I1Cent= 0 AccDes= 0.00D+00.
Defaulting to unpruned grid for atomic number 44.
Defaulting to unpruned grid for atomic number 44.
Defaulting to unpruned grid for atomic number 44.
---------------------------------------
The end of file looks OK, Gaussian is not reporting here any problems. So:
0) Split the optimization and frequency calculations into two separate jobs.
1) if you're running it on a supercomputer with queueing system: perhaps you're running out of walltime for this job?
2) if not: how large is the system (no of atoms?, no of basis functions?) and what method and basis set are you using? If the frequency job is run on a personal computer it can run out of memory/disk space in no time (although Gaussian usually complains about that in the output) due to the enormous requirements. Also, check the size of the .rwf file created in this run.
Dear Dr Bartosz,
Thank you for your reply.
My molecule is of similar size with [Ru(bpy)3]2+ (not big, isnt it) and i used B3PW91 functional/ SDD ECP and coresponding basis set for Ru + 6-31G*
for others.
0, 1: Yes. When it failed, i used the optimized geometry to set up an other freq calculation. This new freq-only should be within the time limit but still failed.
2: I encountered the enormous writing error before, and the error was stated. This time, no error reported. But i still suspect this is the problem, as when clean my disk and re-submit the same opt+freq job, it terminated normally.
I am still monitoring the situation and posted this question to learn experience from people like you and also to share.
Thank you very much, Dr Bartosz.
Regards,
-Mer-
Thank you, Dr Oleg.
I tried to follow your instruction, increased the memory limit then the job terminated normally.
Thank you very much!
-Mer-
After basis set give one space, write 'opt' again give one space then write 'freq'
- Badges
- Science topic