This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [tree-ssa PATCH] Pick memory consumption low hanging fruit
- From: law at redhat dot com
- To: Diego Novillo <dnovillo at redhat dot com>
- Cc: Steven Bosscher <s dot bosscher at student dot tudelft dot nl>, Andrew Macleod <amacleod at redhat dot com>, gcc-patches <gcc-patches at gcc dot gnu dot org>
- Date: Fri, 21 Nov 2003 16:14:36 -0700
- Subject: Re: [tree-ssa PATCH] Pick memory consumption low hanging fruit
- Reply-to: law at redhat dot com
In message <1069450788.5285.24.camel@frodo.toronto.redhat.com>, Diego Novillo w
rites:
>On Fri, 2003-11-21 at 02:21, law@redhat.com wrote:
>> In message <200311200922.02374.s.bosscher@student.tudelft.nl>, Steven Bossc
>her
>> writes:
>> > -O2 -fno-tree-pre idem. + -fno-tree-dominator-o
>pts
>> >Total Allocated 862045732 (+12% ?!) 718380296
>> Could very easily be the additional varrays. I've got a patch which
>> cuts them down significantly. Unfortunately, that change has exposed
>> a latent bug. The bug is in a hunk of code I'm not particularly familiar
>> with, so I've passed it to Diego.
>>
>This ought to fix the bug you found. Got a test case we can add to the
>testsuite?
>
>We were not checking _REF nodes properly to see if their base address
>was in global memory. Will commit after bootstrap/tests are done.
>
>
>Diego.
>
>
> * Makefile.in (tree-dfa.o): Add dependency on $(TREE_DUMP_H)
> * tree-dfa.c: Include tree-dump.h
> (compute_alias_sets): Call dump_function_to_file.
> (may_access_global_mem_p): Check if the base address of _REF nodes
> may point to global memory.
I've done a bootstrap and test with this patch (well, the
may_access_global_mem_p part, which is what's important). I've done some
sanity checking on the rest.
I've got code which is dependent on getting this fixed, so I'm going to
go ahead and check it in.
Jeff