[MITgcm-devel] Re: [MITgcm-support] bug in exch2?

[Dimitris Menemenlis]

Martin, since the exchanges are now done outside scatter_2d.F, i.e., after 
calling MDSREADFIELD, why not simply remove the overlap regions, i.e.,

       _RL     local(1:sNx,1:sNy,nSx,nSy)
       _RL     temp(1:sNx,1:sNy,nSx,nSy)
       lbuff = sNx*nSx*sNy*nSy

in scatter_2d.F and gather_2d.F

       _RL sharedLocalBuf(1:sNx,1:sNy,nSx,nSy)

in MDSIO_SCPU.h, and corresponding changes in mdsio_passtorl.F, etc.?

This will take care of uninitialized edges as well as speed up the code a little 
bit.  If you agree I can go ahead and do changes.

D.

> After my previous premature babbling I have now found a fix to the problem: I
> initialize "local" (which is the name of the argument on which sharedLocalBuf
> is passed) and "temp" in scatter_2d and the nans in the overlaps go away.
> 
> If this is a general solution, I'll happily check this in (but I'll wait for
> your approval, as I do not have enough of an overview on this issue).