This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Structure Return Testcase
- To: Richard Henderson <rth at redhat dot com>
- Subject: Re: Structure Return Testcase
- From: Jeffrey A Law <law at redhat dot com>
- Date: Thu, 15 Mar 2001 16:20:21 -0700
- cc: "Joseph S. Myers" <jsm28 at cam dot ac dot uk>, gcc at gcc dot gnu dot org
- Reply-To: law at redhat dot com
In message <20010315130847.A22538@redhat.com>you write:
> On Thu, Mar 15, 2001 at 01:54:46PM -0700, Jeffrey A Law wrote:
> > Do we think it makes more sense to use memmove or to do the copy through
> > a temporary?
>
> Memmove, definitely. The bits are only shuffled once. Do that
> twice and you double memory bandwidth requirements.
OK. Yea, I guess that makes sense, even given lame implementation of
memmove.
> > Also note Loren's patch pessimizes basic structure copies, which probably
> > is bad.
>
> I don't think so. I thought the same thing until Loren
> pointed out that it was keyed to RESULT_DECL, which is
> by definition a return value.
Ah. OK.
jeff