This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: FW: GNU f77 bug report
- To: bessel at earthlink dot net
- Subject: Re: FW: GNU f77 bug report
- From: craig at jcb-sc dot com
- Date: 27 May 1999 04:46:01 -0000
- Cc: egcs-bugs at egcs dot cygnus dot com
- Cc: craig at jcb-sc dot com
- References: <01BEA7CE.8CF32880.bessel@earthlink.net>
>while (t1 != t1o)
> {...}
>
>where t1 and t1o are real*8.
Then it's most likely a case of assuming convergence to perfect
equality, when that doesn't always happen, especially on machines
like the x86 (most especially given that egcs/gcc don't spill intermediate
FP results with all 80 bits intact).
Look up "Floating-point Errors" in the g77 docs, for pretty much the
whole story, much of which is linked-to from there.
If you can track down an actual code-gen bug *other* than a truncating
spill, though, then by all means let us know!
tq vm, (burley)