[Wien] Error in lapw1para_lapw script causing errors when running parallel lapw2

[Peter Blaha]

Testing this in TiC with 47 k-points:

4 lines in .machines; granularity:1

testpara_lapw   produces:

1 : homer(11) 11k
2 : homer(11) 11k
3 : homer(11) 11k
4 : homer(11) 11k
5 : homer(11) 3k

and also x lapw1 -p / x lapw2 -p runs fine.

It produces 5 !!! klists, the latter one with the remaining 3 k-points and the
"fastest" cpu will get this junk.

So from my point of view it works perfectly well.

 > Indeed I am using $SCRATCH variable. I've also checked -it switch and it
 > works with k-points splitted 2/2/2/2/1  on 4 cpus.

No, you cannot use iterative diag, because with 4 lines in .machines, but
actually 5 !! junks, you don't know on which computer the 5th junk will be executed
(it will be the fastest, but that can change from iteration to iteration and you will
not have an old vector file.

Paweł Leśniak, IFMPAN schrieb:
> W dniu 2009-06-16 23:25, Peter Blaha pisze:
>> Can you send me your    .machines   file
>> (and eventually some options set in $WIENROOT/parallel_options)
>>
> $ cat .machines
> granularity:1
> 1:b03
> 1:b03
> 1:b03
> 1:b03
> 
> $ cat /opt/phys/wien2k_09.1.test/parallel_options
> setenv USE_REMOTE 1
> setenv WIEN_GRANULARITY 1
> 
>> I cannot remember such a problem and we do k-parallelization since 
>> many years. (Although I admit, usually I create my .machines file in a 
>> "meaningful" way).
> That's the point. Sometimes it works ok, but sometimes it's needed to 
> change number of k-points so reduced number of k-points is dividable by 
> "number of cpus". That's because of missing two k-points caused by 
> mistake in line 444th.
> 
> I've checked older versions of WIEN2k  (versions 7th and 8th) and 
> there's the same problem. I've no older versions installed anywhere, so 
> I'm unable to check how old this mistake is.
> 
> I've sent previously klists with original formula and corrected one, 
> where you can see that corrected formula produces correct split of 
> k-points. There's also mistake caused by this corrupted formula in 
> manual. We can read: "When using iterative diagonalization" or the 
> $SCRATCH variable (...), the ratio (k-points / processors must be 
> integer (...)" which is not correct at all - just correction of formula 
> in line 444th is needed.
> 
> Indeed I am using $SCRATCH variable. I've also checked -it switch and it 
> works with k-points splitted 2/2/2/2/1  on 4 cpus.
> 
> Pawel Lesniak
> 
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien

-- 

                                       P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------