Inconsistency in r4 configuration data.exch2

[ecco-group]

Inconsistency in r4 configuration data.exch2

[emmomp]

This seems to be deliberate as it was introduced in a commit by @owang01 a9a38c0

Am I missing something else? When I try to run as is (without 188, 242 and 376) I get the error "Domain Total # of tiles = 468 does not match (SIZE.h+blankList)= 465"

[owang01]

@emmomp The files in this repository will be kept being updated, while those hosted by the PO.DAAC (or the NASA drive as you called) are frozen. So when there is difference, use files in this repository.

This version of data.exch2 (without 188, 242 and 376) is deliberate. @jahn pointed out that with the three tiles he got unconnected-points warnings and the results diverge near those unconnected-points. The pull request #105 (commit ommit a9a38c0) reflects this change.

Regarding the error "Domain Total # of tiles = 468 does not match (SIZE.h+blankList)= 465", I suspect the variable "nPx" is still set to 360 in your SIZE.h. If so, please change it to 363 and recompile the code. I believe this will fix the error.

I just submitted a new pull request #152 that changes "blankList(1:108)" in data.exch2 to "blankList(1:105)". However, this bug is benign. The model states before and after the fix are identical.

[emmomp]

Thanks for clarifying @owang01 I was getting Nans for my 360 setup so that might be the reason.

For clarity to users, it might be worth also changing "nprocs=360" to "nprocs=363" and/or providing a SIZE.h.363 file that is consistent with data.exch2.