parallel distribution of data

[Matt W]

OK,  this last report shows it is almost certainly an overflow at

routine pair2int (realspace_task_selection.F, line 72)

I can't run this system here, but Teo if you can could you check
whether the result of realspace_task_selection.F line 72 is ever
negative (or any other better test for an overflow)?

The memory use should not be larger than replicated!

Matt

On Mar 10, 9:53 pm, Teodoro Laino <teodor... at gmail.com> wrote:
> Ciao Nick,
>
> WOW.. that's huge.. 1.9 GB per procs..
> I will try to run it but I'm hopeless, if this is the error..
>
> I guess this is something that Matthew should look into..
> Anyway will let you know once I have the result..
>
> Cheers
> teo
>
> On 10 Mar 2008, at 22:49, Nichols A. Romero wrote:
>
> > *** The memory allocation for the data object <send_buf_r> failed.
> > The  ***
> >  *** requested memory size is 1931215
> > Kbytes                             ***