This is the mail archive of the
libc-alpha@sources.redhat.com
mailing list for the glibc project.
Re: PATCH: Optimize memcmp for ia32
On Fri, Feb 06, 2004 at 11:27:30PM -0800, Ulrich Drepper wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> H. J. Lu wrote:
>
> > I tested it on all available Pentium 4/M processors.
>
> And because of your tests with P4s you replace the code for generic
> i386? Not very logical, is it?
I don't think anyone will be happy to run the current glibc on anything
less than Pentium Pro.
>
> Furthermore, the test case code is not a realistic benchmark. For this
> you need to know the usual string sizes which happen on a normal system.
> Do you have this data? Well, I have some code which does measurements
> of the use of all string functions. And guess what the outcome is?
> memcmp in the libc is not used at all during a normal system start and
> shutdown. I've even started OO.o.
>
The patch is to improve the speed of memset in glibc. Programs which
use memset in glibc will benefit from it. One of them is gcc:
# nm -D /usr/lib/gcc-lib/i386-redhat-linux/3.2.2/cc1| grep memset
U memset
We can investigate how often it is called in gcc. But it is used by
some applications and my patch will help them.
H.J.