[MITgcm-devel] Re: notice the last checkin?

[Chris Hill]

Unfortunately I think Ed is right i.e. endianess also applies to integers so
the implied do loop would have a problem with g77. Apart from netCDF other
options are

 (1) the PGI compiler allows you request byte swapping I/O.  
 (2) Matlab can be used to swap bytes offline. 


Chris 
 

> -----Original Message-----
> From: mitgcm-devel-bounces at mitgcm.org 
> [mailto:mitgcm-devel-bounces at mitgcm.org] On Behalf Of Ed Hill
> Sent: Tuesday, July 20, 2004 12:53 PM
> To: MITgcm-devel
> Subject: [MITgcm-devel] Re: notice the last checkin?
> 
> On Tue, 2004-07-20 at 12:27, Andrea Molod wrote:
> > hi ed,
> > just wanted to make sure you took notice of the last 
> checkin i did for 
> > fizhi. it now does the byteswapping if it needs to. i 
> forget whether 
> > its big_endian or little_endian, but i know that if you 
> make a file on 
> > a dell you need to do the byteswapping before writing to 
> get it all to 
> > come out ok. AND that the files you will read from my data/bcs 
> > directory were created on an SGI so need to be byteswapped 
> to be read 
> > on the linux box.
> 
> 
> Hi Andrea,
> 
> I'm moving this over to the devel list since I'd like Chris 
> to see it and am hoping he will correct me if I'm wrong.  [Please!]
> 
> The byte-swapping needs to be done for all variables that 
> have a size greater than one byte.  Thus, code such as:
> 
> ====== pkg/fizhi/update_ocean_exports.F : Lines 653--661 ======
> 
>       read(iunit,rec=nrec,err=500) cname, cdscrip, 
>      .     im_32, jm_32, npx_32, npy_32, lat0_32, lon0_32, 
>      .     ndatebc_32, undef_32,
>      .     (nymdbc_32(i), nhmsbc_32(i), i=1,ndatebc_32)
> #ifdef _BYTESWAPIO
>       call MDS_BYTESWAPR4( 1, lat0_32)
>       call MDS_BYTESWAPR4( 1, lon0_32)
>       call MDS_BYTESWAPR4( 1, undef_32)
> #endif
> 
> ====== pkg/fizhi/update_ocean_exports.F : Lines 653--661 ======
> 
> will not work with the same input files on both big- and 
> little-endian systems because all the variables (including 
> the "integer*4 ndatebc_32" 
> in the implied read loop) must be byte-swapped before they 
> can be used.
> 
> This byte-swapping problem is one of the many advantages of 
> NetCDF since all I/O done through calls to NetCDF routines 
> are (transparently to the
> caller) correctly byte-swapped.
> 
> Ed
> 
> --
> Edward H. Hill III, PhD
> office:  MIT Dept. of EAPS;  Rm 54-1424;  77 Massachusetts Ave.
>              Cambridge, MA 02139-4307
> emails:  eh3 at mit.edu                ed at eh3.com
> URLs:    http://web.mit.edu/eh3/    http://eh3.com/
> phone:   617-253-0098
> fax:     617-253-4464
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://dev.mitgcm.org/mailman/listinfo/mitgcm-devel
>