[apt4ssx8] Re: Compiling apt360 with g77
- From: Brent Muller <bmuller@xxxxxxxxxxxx>
- To: apt4ssx8@xxxxxxxxxxxxx
- Date: Wed, 27 Jun 2012 13:13:05 -0600
On 6/27/2012 10:54 AM, Jon Elson wrote:
Brent Muller wrote:
There are two things that come to mind when porting: one, what would
be the affect of running code made for big-endian machines on intel?
Properly written pure Fortran code should be very unlikely to suffer
problems with this.
You could certainly run into issues with assembler routines, or in
places where
several byte variables are commoned to a wider variable. Otherwise,
the byte order
I think this problem, where smaller byte variables are commoned into a
wider variable, was precisely the problem. At least that's how the C
code of f2c generated it. To see the areas in the C code that you might
cross-reference with the Fortran to verify, grep for "OTHER_ENDIAN" macros.
in a 32-bit variable should always be the same, so you should be able
to ignore
the endian-ness, as long as all I/O is character-based. If you read
or write
binary files, then this becomes a big issue.
This shouldn't be a problem, since binary file writing was handled by
assembler routines
I overcame this problem in the C code with macros everywhere this
was a problem. Two, how do we call C code in the ASM file conversion
from Fortran?
Mostly, it "just works". The one place where there is a big
difference is that the way
two-dimensional arrays are laid out in memory are opposite for C and
Fortran.
Jon
Other related posts:
