-O2 on alphaev6 has big problems
Brad Lucier
lucier@math.purdue.edu
Fri Nov 10 11:01:00 GMT 2000
> From: Geoff Keating <geoffk@geoffk.org>
>
> Brad Lucier <lucier@math.purdue.edu> writes:
>
> > What is somewhat frustrating is that while I can't get the CVS sources
> > to compile my code correctly at all, I see the results on the test
> > suite getting better and better.
>
> This is a sign you aren't reporting the bugs you find. If you
> reported bugs, with a small example suitable for the testsuite, they
> would go into the testsuite, and then you would see the failure count
> increase. Eventually, someone would fix the failure and suddenly your
> code would work :-).
I don't know how to interpret this comment. There are 124 hits of "Lucier"
in gcc-bugs; the report I referred to in my original post,
http://gcc.gnu.org/ml/gcc-prs/2000-q4/msg00208.html
shows exactly where the bad code is inserted. I just don't know how to
use gdb on cc1 to break precisely when gcse attempts to add this code
to this block to find out *why* it wants to do this.
Brad
More information about the Gcc
mailing list