The ABINIT Users Mailing List ( CLOSED )

[Xavier Gonze <gonze@pcpm.ucl.ac.be>]

Dear Andriy,

Andriy Nevidomskyy wrote:
> Dear Developers of ABINIT,
>
> I have the following question to ask. What is the present status of the
> parallelization over bands in ABINIT?
>
>        The 'Features' file gives the following info:
> "B.1. For ground-state calculations, the code has been parallelized 
> on the k-points, on the spins, on the bands,
> and on the FFT grid and plane wave coefficients."
>  From what is written above one would assume that parallelizations of all
> kinds should work equally well (with MPI).
>             However:
>  I was running the parallel version of ABINIT_3.4.3 on IBM-SP3 machine,
> but whatever I tried, the band parallelization wouldn't work. 
> For example, in the case of 1 k-point
> all the load is carried by only ONE processor, while the others are idle!
>
> I used the following parameters in the input file, as documented in
> the list of input variables:
>    wfoptalg 1     # blocked CG algorithm minimization 
>    nbdblock 8     # here, say, 8 - number of available processors

This seems correct indeed ...

> I played around with different values of <nbdblock>, being integer
> divisors
> of the total number of processors, but it had no effect on the run. 
> In fact, the sequential version of ABINIT did better job in all cases!

So, using "abinis" with the same input file than "abinip" was faster ?

> Am I missing something essential, which is why the parallel version
> doesn't work, or is there any other reason for this?
> I would be very grateful for your help.

At present, I do not see where is the problem. Thierry Deutsch
did tests of the band parallelisation some months ago,
and things seemed to be OK... You should observe a difference !

> Yours Sincerely,
> Andriy Nevidomskyy
>
> P.S> The parallelization over k-points seems to be working fine. 
> Unfortunately only when nkpt=integer*number_of_processors.

?! It works also if nkpt is not a multiple of the number of processors.
Simply, the load balancing is not as good, so there is some (small)
lost in efficiency.

So, there must be some problem, but where ?
Xavier