From:	khan AT xraylith DOT wisc DOT edu (Mumit Khan)
Subject:	Re: g77
2 Feb 1997 13:10:43 -0800 :
Approved:	cygnus DOT gnu-win32 AT cygnus DOT com
Distribution:	cygnus
Message-ID:	<9702021912.AA05976.cygnus.gnu-win32@modi.xraylith.wisc.edu>
Original-To:	"Michael Hung" <hungm AT hotmail DOT com>
Original-Cc:	gnu-win32 AT cygnus DOT com
In-Reply-To:	Your message of "Sat, 01 Feb 1997 08:20:24 PST."
	<199702011620 DOT IAA21168 AT f16 DOT hotmail DOT com>
Original-Sender:	owner-gnu-win32 AT cygnus DOT com

"Michael Hung" <hungm AT hotmail DOT com> writes:
> Hello, is there a g77 for use with cygnus-win32 ?
> 
> Thanks
> Michael
> 

I've sent the patches to gnu-win32 to g77 maintainers for both b16 and
b17 (there was a minor interface change in gcc). In B16, there was a
known bug in numerics when optimization was turned on, and looks like
it's fixed in b17. The big remaining problem is performance of
unformatted I/O, and I've had to write C routines to read/write
unformatted files for our code (not a general solution obviously). The
problem is due to the f2c's method of fseek'ing to write the record
length and the corresponding flushing of the file descriptors kill the
performance (something like 10 times slower on large unformatted files
with small records).

If you're interested, email and I'll email/post the diffs (very small)
and my comments (much larger than the diffs :-). btw, I've built g77
only as a x-compiler on a linux box, so mileage will definitely vary on
a native build.

Regards,
Mumit -- khan AT xraylith DOT wisc DOT edu
http://www.xraylith.wisc.edu/~khan/
-
For help on using this list, send a message to
"gnu-win32-request AT cygnus DOT com" with one line of text: "help".

- Raw text -