The ABINIT Users Mailing List ( CLOSED )

["Guillaume Dumont" <dumont.guillaume@gmail.com>]

Dear Dr. Bottin,

What do you mean by // ???

The orthon in vtowfk.F90 was removed in //

Does this problem appear for all the distributions (npband;npfft)? 

I'm going to check this for my GaAsN supercell case, but see below

Does this problem exist before the 2 SCF? Is there any swap in your calculations
of speedup?

Yes. Not for the GaAsN case but for a Metal-organic framework with a C60 in it... 

nband 504
natom 226
ngfft 272 272 272

For some number of processors the code does not complete a single SCF, for some other in does one, and some other it does 2. This also depends on the set of npband and npfft. Not surprising since the memory need seems to depend this as well. For example, for 16 procs run if

npband=npfft=4 we get 2 SCF cycles

but

if

npband=8 and npfft=2 we get 1 SCF.

Thanks for your time.

On 11/30/06, Francois Bottin < Francois.Bottin@cea.fr> wrote:

Dear Dr. Dumont,
This is the orthon routine which permits to start with orthogonalized
wavefunctions. The orthon in vtowfk.F90 was removed in //. These
routines are very time consuming when nprocs or nband increase. You
could remove this one in // as it will be in the 5.3.x release. Note
that these two orthon have no influence on our results (with the lobpcg
algorithm).
Regards
Francois Bottin

Guillaume Dumont wrote:

> Hi,
>
> I have noticed in my output files (for the gold case) that when the
> number of processors is increased the inwffil routine starts to become
> very time consuming, around 30% of total time at 144cpus. Any idea
> what could cause this routine to take that much time?
>
> regards


--
##############################################################
Francois Bottin                    tel: 01 69 26 41 73
CEA/DIF                            fax: 01 69 26 70 77
BP 12 Bruyeres-le-Chatel         email: Francois.Bottin@cea.fr
##############################################################

-- 
Guillaume Dumont
=========================
guillaume.dumont.1@umontreal.ca
dumont.guillaume@gmail.com
(514) 341 5298
(514) 343 6111 ext. 13279