[Pw_forum] parallel diag. "failure" with too large number of k-points?????

[Paolo Giannozzi]

Giovanni Cantele wrote:

> The calculation is parallel and uses 8 CPUs

with or without k-point parallelization?

> All the grids with a number of k-points > 804 show a strange behaviour, 
> namely, that once reached the 804th k-point
> (so I see in the output the "Computing kpt #:   804" line) the 
> calculation remains running but no update of output files is done.
> 
> Has anybody experienced such kind of problem?

of course, no. Is this reproducible on different machines?

> the one with "-ndiag 1" runs much faster (not sure but maybe half the time)

subspace diagonalization in a calculation like yours should take a small
fraction of the total. Unless you have a really crappy machine, I don't
see how this small fraction of time can grow to 50%

P.
-- 
Paolo Giannozzi, Democritos and University of Udine, Italy