[molpro-user] Failing testjobs loc_eom.test and n_uhf_fciqmc.test
Tuononen Heikki
heikki.m.tuononen at jyu.fi
Wed Nov 5 11:00:29 GMT 2014
Hi,
I just compiled Molpro 2012.1 with ifort (version 14.0.3) and mkl (from composer_xe_2013_sp1.3.174) in serial mode and then run all the testjobs. There seems to be two with problems:
The first:
ERRORS DETECTED in loc_eom.test. Max error for energy=1.00570617D-07
STATE ENERG ENERG-ENERGYOLD
1.1 -154.5350562 -1.71978343D-08
2.1 -154.1221469 -1.15337002D-08
3.1 -154.0870278 -1.90899812D-08
4.1 -154.0860293 -2.65540336D-08
5.1 -154.0322256 -2.63903530D-08
1.1 -154.5346723 -2.06508446D-08
2.1 -154.1112551 -1.57624243D-08
3.1 -154.0841127 -3.78332459D-08
4.1 -154.0807804 -1.57382374D-08
5.1 -154.0202992 -2.70004250D-08
1.1 -154.5346723 -2.06508446D-08
3.1 -154.0872525 -1.69885368D-08
1.1 -154.5346723 -2.06508446D-08
3.1 -154.0754928 -5.29949489D-08
1.1 -154.5346723 -2.06508446D-08
2.1 -154.1297802 -5.80482151D-09
3.1 -154.0920520 -2.56981991D-08
1.1 -154.5346723 -2.06508446D-08
2.1 -154.1216660 -1.26034365D-08
3.1 -154.0860756 -1.62419838D-08
1.1 -154.5351216 -1.75655543D-08
3.1 -154.0853344 -2.88179365D-08
1.1 -154.5351216 -1.75655543D-08
3.1 -154.0765951 -6.59164243D-08
2.1 -153.8431778 -3.88002945D-08
3.1 -153.8286514 -1.71936847D-08
1.1 -154.5346723 -2.06508446D-08
3.1 -154.0725698 -1.00570617D-07
1.1 -154.5346723 -2.06508446D-08
3.1 -154.0725698 -1.00570617D-07
1.1 -154.5346723 -2.06508446D-08
3.1 -154.0725698 -1.00570617D-07
1.1 -154.5346720 -1.40349243D-08
2.1 -154.0844799 -2.26346231D-08
1.1 -154.5346720 -1.40349243D-08
2.1 -154.0844799 -2.26346231D-08
The result for the state 3.1 (-154.725697706) is slightly lower than the reference value (last three cases).
The second:
ERRORS DETECTED in n_uhf_fciqmc.test. Max error in energy=2.17459737D-05
PROG ETOT ERRTOT ETOT-EOLD ERRTOT-ERROLD
UHF -54.39111456 0.00000000D+00 -2.13277218D-09 0.00000000D+00
FCIQMC -54.48011121 2.68997785D-05 2.17459737D-05 6.10977845D-06
FCI -54.48011505 0.00000000D+00 -4.24677182D-09 0.00000000D+00
This seems to be relevant to bug 3784. However, I thought that this is fixed for serial jobs already to give consistent results across all compilers and platforms.
Best wishes,
Heikki
Dr. Heikki M. Tuononen
Department of Chemistry
University of Jyväskylä
Finland
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20141105/4c13e74f/attachment.html>
More information about the Molpro-user
mailing list