This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [tree-ssa] Generic NRV optimization [was Re: two-element struct performance (was: strict-aliasing and typedefs) ]
- From: law at redhat dot com
- To: Diego Novillo <dnovillo at redhat dot com>
- Cc: Andreas Schwab <schwab at suse dot de>, Joe Buck <jbuck at synopsys dot com>, Gabriel Dos Reis <gdr at integrable-solutions dot net>, Brad Lucier <lucier at math dot purdue dot edu>, "gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>
- Date: Thu, 26 Feb 2004 11:14:41 -0700
- Subject: Re: [tree-ssa] Generic NRV optimization [was Re: two-element struct performance (was: strict-aliasing and typedefs) ]
- Reply-to: law at redhat dot com
In message <1077718049.16194.206.camel@localhost.localdomain>, Diego Novillo wr
ites:
>On Wed, 2004-02-25 at 08:47, Andreas Schwab wrote:
>> law@redhat.com writes:
>>
>> > Bootstrapped and regression tested on i686-pc-linux-gnu.
>> >
>> > * Makefile.in (OBJS-common): Add tree-nrv.o.
>> > (tree-nrv.o): Add dependencies.
>> > timevar.def (TV_TREE_NRV): New timevar.
>> > * tree-nrv.c: New file implementing NRV on generic trees.
>> > * tree-optimize.c (init_tree_optimization_passes): Link in
>> > tree_nrv optimization pass.
>> > * tree-pass.h (tree_nrv): Declare.
>> > * doc/invoke.texi: Add -fdump-tree-nrv documentation.
>>
>> This breakes bootstrap on ia64-linux.
>>
>> stage1/xgcc -Bstage1/ -B/usr/local/ia64-suse-linux/bin/ -c -g -O2 -DIN_G
>CC -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -pedan
>tic -Wno-long-long -Wold-style-definition -Wno-variadic-macros -Werror -fno-c
>ommon -DHAVE_CONFIG_H -I. -I. -I../../gcc -I../../gcc/. -I../../gcc/../i
>nclude -I../../gcc/../libbanshee/libcompat -I../../gcc/../libbanshee -I../../
>gcc/../libbanshee/points-to ../../gcc/tree-ssa-ccp.c -o tree-ssa-ccp.o
>> ../../gcc/tree-ssa-ccp.c: In function `get_default_value':
>> ../../gcc/tree-ssa-ccp.c:2028: internal compiler error: in expand_expr_real
>_1, at expr.c:6652
>>
>ia64 was broken already, but this is the same ICE we are getting on
>x86-64.
>
>Jeff, we get confused in change_address_1 because we feed it a PARALLEL
>rtx
I just got back into town. I'm looking at it now...
jeff