[MITgcm-devel] coordinate variables

Jean-Michel Campin jmc at ocean.mit.edu
Sat Dec 18 14:03:23 EST 2004


Hello Ed,

Can you comment out the writing of stuff like:
 bi,bj =  1, 1   dname = T
to standard output file. (I only get ~100 of them, but I am 
not writing many mnc files).
I don't think it' really usefull for anyone except you.

Also, in mnc_cw_cvars.F, I suggested to have rtmp declared  
      _RS rtmp(sNx + sNy)
instead of
      _RS rtmp(sNx * sNy)
because (I said) "it will take less memory".
Now I still think my suggestion is worth to implement, but for
a different reason:
In a 2 D configuration, sNx * sNy is not enough, but sNx + sNy
works well:
I tried mnc with ideal_2D_oce with ifc & check-bound and I get
> Run-Time Error 406: Array bounds exceeded
>    In Procedure: mnc_cw_write_cvar

And finally, what happens to those coordinate variables if
I want to write "with full overlap" instead ?
Is this option still available ?

Jean-Michel



More information about the MITgcm-devel mailing list