Search found 4 matches

by Artem
Fri Oct 26, 2018 10:14 am
Forum: Ground state
Topic: large SCF iteration time: wrong paralellism parameters?
Replies: 3
Views: 3169

Re: large SCF iteration time: wrong paralellism parameters?

Dear Boris, Thank you very much for your reply. I gave a try to subsitute autoparal=1 with manual setting and increased number of empty bands instead of default 5% of total VB by recomended 192 to keept it even, but it yielded worse results: BINIT 8.6.1 Give name for formatted input file: aba1206171...
by Artem
Thu Oct 18, 2018 9:35 am
Forum: Ground state
Topic: large SCF iteration time: wrong paralellism parameters?
Replies: 3
Views: 3169

Re: large SCF iteration time: wrong paralellism parameters?

There is some update on my recent post: overnight run afforded 5 SCF iterations, each took from 2-4 hrs. ---SELF-CONSISTENT-FIELD CONVERGENCE-------------------------------------------- getcut: wavevector= 0.0000 0.0000 0.0000 ngfft= 120 144 72 ecut(hartree)= 30.000 => boxcut(ratio)= 1.93937 ITER ST...
by Artem
Wed Oct 17, 2018 6:07 pm
Forum: Ground state
Topic: large SCF iteration time: wrong paralellism parameters?
Replies: 3
Views: 3169

large SCF iteration time: wrong paralellism parameters?

Hello Everybody! I'm trying to understand why my calculations are so prohibitively slow: 1 SCF iteration takes time to over in hours(!!!). The task is running on 8-processor node, so initially I have used autoparal=1 switched on, however it did help. My system is small molecule semiconductor with pr...
by Artem
Fri Oct 14, 2016 1:39 pm
Forum: Link to other codes
Topic: problem with ONCV PS for H
Replies: 1
Views: 4034

problem with ONCV PS for H

Dear Abinit Users! I have faced with a problem using tabulated atomic ONCVPSP data for Hydrogen from current Pseudo Dojo. Simple optimization run aborts with a message: --- !ERROR src_file: psp8in.F90 src_line: 423 mpi_rank: 0 message: | Bad integer for item 1 in list input And following service inf...