| 6:53:09 AM PDT - Sun, Aug 25th 2019 |
|
Actually, the calculation gets stuck at
CCSD iterations
-----------------------------------------------------------------
Iter Residuum Correlation Cpu Wall V2*C2
-----------------------------------------------------------------
When HF is deleted and io ga is added, it still gets stuck.
QA test tce_c2_cc2 also gets stuck.
I arbitrarily choose QA tests tce_n2, tce_cr_eom_t_ch_rohf, tce_eomccsd_gradient, tce_eomsd_somsol1,tce_eomsd_eomsol2,tce_lr_ccsd_tq, tce_c20_triplet, tce_hyperpolar_ccsd_small, tce_helium_imaginary and tce_eaccsd_ozone to eaxamine whether TCE can still work on my computer without the recompilation of NWChem6.8, fortunately all being able to finish. Thus, perhaps there is a bug in cc2 of TCE.
QA test tce_c20_triplet has many warnings during ROHF iterations, diffferent from that in the test.
QA test tce_hyperpolar_ccsd_smallgets stuck when running paralelly.
The output of QA test tce_helium_imaginary is not provided, and I do not know the physical meaning of the calculation, also whether it is correct.
QA test tce_rohf_uccsd can pass, and when I modify it into ccsd(t), it also can pass.
I have reported these and put all the log files on Github.
By the way, in order to obtain osccilator strengths, one should do left-hand eom-ccsd, but it seems NWChem only can do right-hand iterations, whereas GAMESS has this capability.
|
|
|