[MITgcm-devel] testing on ACES

[Jean-Michel Campin]

Hi Greg,

seems to be better today: it took ~1.h for the remove and 
an other ~1.h for the check-out, and we have already few tests
(3/4: g77, ifort & ifort-mth) completed (and OK).
Thanks,
Jean-Michel

On Sun, Apr 08, 2007 at 04:30:54PM -0400, Jean-Michel Campin wrote:
> Hi,
> 
> testing on ACES is not working so well:
> - had a disk full on /net/ds-01/scratch-5 last Friday (Apr 06)
> - yesterday and today, disk acces to /net/ds-01/scratch-5
>  is very slow. To give an idea, on faulks, working on
>  a local disk, it takes less than 8 mn to remove the previous test
>  dir and to download a fresh copy of the code. On the computer nodes
>  of ao, it takes ~5.h to remove the previous dir (even if they contains
>  much less), and after an other 10.h, the downloading (using cvs) is
>  far from being over.
>  The consequence is that after 24.h, only few test (at best)
>  will be completed, and it will stop by reaching the PBS job
>  time limit.
> suggestions ?
> 
> Jean-Michel
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel