This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patch] i386.c string alignment - no alignment for < 32 bytes, else to 32 bytes
- From: Chris Sears <cbsears_sf at yahoo dot com>
- To: John Wehle <john at feith dot com>
- Cc: gcc-patches at gcc dot gnu dot org, gcc at gcc dot gnu dot org
- Date: Tue, 15 Jan 2002 17:46:56 -0800 (PST)
- Subject: Re: [patch] i386.c string alignment - no alignment for < 32 bytes, else to 32 bytes
John
unfortunately, I'm in license limbo with the Intel C++ compiler.
Tomorrow I will compile with it and find out what it does.
I assume since you referred to the IA Optimization manual, you would
accept their compiler as a piece of evidence. Intel documentation,
since there are AMD's etc out there, has its fair share of mistakes
and false leads. Better to see what Intel does rather than says.
> > Let's take the second case first. Applying this rule to *all*
> 32-byte
> > strings is trading off a trivial amount of optimization for memory.
> > These constants are not dynamically allocated objects, they are
> static
> > constants rarely used.
>
> It depends on the application.
I'm sure there is an exception but I'm straining to imagine it.
Remember declarations are handled seperately.
I think you are mistaken here and it doesn't depend on the application.
It is just wasted space with no performance benefit.
> > It is just wasted space.
>
> -Os says to favor space for time.
-Os doesn't affect this.
Chris Sears
cbsears_sf@yahoo.com
__________________________________________________
Do You Yahoo!?
Send FREE video emails in Yahoo! Mail!
http://promo.yahoo.com/videomail/