CCL: Re: RESTARTing a CCSD(T) job....

[Robert Swofford <swofford[at]wfu.edu>]

 Thanks to several who suggested adding SCF=RESTART to the route.
 Unfortunately, G98 does not just pick up a CCDS(T) single-point energy
 where it left off. The final lines of the logfile are:
  (Enter /opt/gaussian98-a.11.4/g98/l401.exe)
  SCF N**3 symmetry information disabled.
  Initial guess natural orbitals from previous density.
 Error retrieving densities from rwf 20603 LenDen=     22052 but
 Length=     11027.
  Error termination via Lnk1e in /opt/gaussian98-a.11.4/g98/l401.exe.
  Job cpu time:  0 days  0 hours  0 minutes  6.6 seconds.
  File lengths (MBytes):  RWF=    6 Int=    0 D2E=    0 Chk=    7 Scr=   1
So just having the .chk file is not enough. The impression is that a
 restart of such a single-point energy calc *might* work if all the rwf
 files were still available (all 11Gb of them!), but alas, the job
 script
 politely cleans up after itself and deletes those.
The only solution I could come up with was just to run the job again,
 using the .chk from the stopped job. The new job then had to run
 through
 all 25-30 iterations of CCSD(T) calc again. Oh, well.
 Bob
 --
 Robert L. Swofford
 Professor of Chemistry
 Wake Forest University
 Winston-Salem, NC 27109-7486
 swofford[at]wfu.edu
 http://www.wfu.edu/~swofford/
 (336)-758-4490 Telephone
 (336)-758-4656 FAX