[MITgcm-devel] darwin_ia32_gfortran

[Martin Losch]

Hi Jean-Michel,

currently there is no regular test with this opt file (Hezi Gildor had one, but that has ceases to work and I do not have access to his machines).

So I'll adapt this file to my machine (which also has some non-standard paths) and check the modifications in. But because I am pressed for time, I'll do a quick fix now (that works for me) and then we'll see what happens.

Martin

On Sep 20, 2011, at 4:56 PM, Jean-Michel Campin wrote:

> Hi Martin,
> 
> I fixed the FFLAGS. 
> And regarding INCLUDES & LIBS setting,
> 1) I think that this is somehow consistent with what
>  was there in the previous optfile & in darwin_ia32_gfortran+mpi
> 2) since I was not able to test this optfile, and don't know
>  how standard those path are (especially in the +mpi one)
>  I did not want to add some problems with netcdf path.
>  So, yes, I was aware of NETCDF_ROOT beeing ignored when using
>  default MPI (without MPI_INC_DIR).
> If you can test improve this (and better, if you have a
> way to test it), please go for it.
> 
> Cheers,
> Jean-Michel
> 
> Oe Mon, Sep 19, 2011 at 03:17:30PM +0200, Martin Losch wrote:
>> Jean-Michel,
>> 
>> just to make sure before I change this: the current lines 52-54 in darwin_ia32_gfortran:
>>> FFLAGS="$FFLAGS -fsecond-underscore -fconvert=big-endian"
>>> #- might want to use '-fdefault-real-8' for fizhi pkg:
>>> #FFLAGS="$FFLAGS -fdefault-real-8"
>> are pretty much useless as FFLAGS is overwritten in the "if tIEEE; then" block. I suggest moving the below this block (i.e. after line 64).
>> 
>> The same is true for the includes and libs that are overwritten by some mpi-paths at the end of the file.
>> 
>> Martin
>> 
>> _______________________________________________
>> MITgcm-devel mailing list
>> MITgcm-devel at mitgcm.org
>> http://mitgcm.org/mailman/listinfo/mitgcm-devel
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel