This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug optimization/10469] [3.3, 3.4] constant V4SF loads get moved inside loop
- From: "rth at gcc dot gnu dot org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 18 Jun 2003 00:38:15 -0000
- Subject: [Bug optimization/10469] [3.3, 3.4] constant V4SF loads get moved inside loop
- References: <20030423204600.10469.rguenth@tat.physik.uni-tuebingen.de>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10469
rth@gcc.gnu.org changed:
What |Removed |Added
----------------------------------------------------------------------------
AssignedTo|unassigned@gcc.gnu.org |matz@suse.de
------- Additional Comments From rth@gcc.gnu.org 2003-06-18 00:38 -------
At least on x86, the load within the loop comes from reload. The pseudo that
holds the V4SF temporary (right from the beginning mind!) gets forced to memory
because of the subregs; we then reload the pseudo as necessary, which results
in the load within the loop.
With -fnew-ra, we get an ICE:
zz.c:19: internal compiler error: in subreg_hard_regno, at emit-rtl.c:1087
Assigning the bug to Michael, since the current problem can't be solved without
the new register allocator, and because of the ICE.