[MITgcm-devel] exf interp changes

[Martin Losch]

Hi Jean-Michel,
happy new year to you, too.

the new exf_interp.F code is fine with me. Vectorization is nearly perfect. Go ahead and remove the old code.

Martin

On Jan 1, 2012, at 2:29 AM, Jean-Michel Campin wrote:

> Hi Martin,
> 
> I made some changes in exf_interp.F, mainly the way to search for
> latitude index, but these modifications should not affect the results.
> I kept the old code (OLD_EXF_INTERP_LAT_INDEX) but this CPP flag
> is undef at the top of exf_interp.F.
> 
> If you could check that it vectorise well on your vector machine,
> it would be nice, so that I will be able to remove the old code.
> 
> The reason is that I would like to make a new S/R exf_interp_uv
> to interpolate the 2 vector components at the same time.
> Oliver show me a plot of the interpolates wind field near the North
> pole, and I think it can be improved if we have the 2 components
> available when we interpolate (so that the wind @ the N.pole
> would be computed by averaging the 2 components of the wind).
> And since this new S/R will be mostly a copy of exf_interp.F,
> if we can remove the old code before, would be simpler.
> 
> Cheers,
> Jean-Michel
> 
> PS: realise that on European time, the year is over.
>    so, Best wishes for 2012 !
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel