--- Additional Comments From pinskia at gcc dot gnu dot org 2005-02-10
19:40 ---
Fixed, thanks zdenek.
--
What|Removed |Added
Status|NEW
--- Additional Comments From steven at gcc dot gnu dot org 2005-02-01
00:33 ---
Patch posted for review for inclusion in GCC 4.0 is here:
http://gcc.gnu.org/ml/gcc-patches/2005-01/msg02207.html.
--
What|Removed |Added
---
--- Additional Comments From rakdver at gcc dot gnu dot org 2005-01-30
11:59 ---
Patch #7 of http://gcc.gnu.org/ml/gcc-patches/2004-12/msg01381.html more or
less solves the problem (there is one small bit missing that does not affect
this testcase, but might be significant elsewhere).
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-01-30
01:55 ---
(In reply to comment #4)
> Also happens on x86-64, but not on x86.
Well the IV selection on x86 is not good either.
*(long long unsigned int *) d.18 = 0;
D.1187 = (long long unsigned int *) ivtmp.4 + (
--- Additional Comments From steven at gcc dot gnu dot org 2005-01-30
01:33 ---
Also happens on x86-64, but not on x86.
--
What|Removed |Added
CC|
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-01-30
01:12 ---
Here is the .optimizate dump for -fno-ivopts (the most optimial code):
{
:
if (w > 63) goto ; else goto ;
:;
*(long long unsigned int *) d = 0;
*(long long unsigned int *) (d + 8B) = 0;
*(long long
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-01-30
01:08 ---
Confirmed, also happens on powerpc-darwin (This might be the one of the reasons
why SPEC is slower
on the mainline on some tests with -fivopts turned on).
Yes this is one of the few true iv-opts problems.
--- Additional Comments From steven at gcc dot gnu dot org 2005-01-30
00:55 ---
Can you add a tree dump and the output with -fno-ivopts?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19701