[mpas-developers] plans for parallel i/o?
Michael Duda
duda at ucar.edu
Thu Nov 10 10:19:39 MST 2011
Hi, Mark.
I've been experimenting with the PIO library that is used in CESM, and I
think a first-cut implementation should be pretty straightforward; the
biggest issue may be how we make the transition to PIO in the trunk.
After we commit our non-hydrostatic code to the trunk (hopefully later
today!), I'm planning to refresh the 'pio' branch that I created a
couple of weeks ago; if all goes well, that branch should have
functional parallel I/O by early next week, at which time we can begin
to use it in a testing mode.
Cheers,
Michael
On Thu, Nov 10, 2011 at 09:17:13AM -0700, Mark Petersen wrote:
> MPAS developers,
>
> The ocean model is now in a state where we are testing high resolution
> simulations. It appears that we are limited by the lack of parallel
> i/o. I think I've seen a few commits on this. Micheal, are there plans
> for parallel i/o for the mpas framework?
>
> When I run with our 15km ocean mesh, which has
> nCells = 1848103 ;
> nEdges = 5578724 ;
> I get an 'insufficient virtual memory' error at the time of the first
> output write.
>
> Thanks,
> Mark
> _______________________________________________
> mpas-developers mailing list
> mpas-developers at mailman.ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/mpas-developers
More information about the mpas-developers
mailing list