This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Haney's real matrix test regression
- To: Paolo Carlini <pcarlini at unitus dot it>
- Subject: Re: Haney's real matrix test regression
- From: pete at ltoi dot iap dot physik dot tu-darmstadt dot de
- Date: Sat, 27 Oct 2001 13:23:02 +0200 (MEST)
- Cc: Richard Kenner <kenner at vlsi1 dot ultra dot nyu dot edu>, <gcc at gcc dot gnu dot org>, <CGLeggett at lbl dot gov>
Hello ...
yesterday i was not here in the office / at my E-mail account;
and i haven't read the gcc-mailing list yet. Thus i currently don't
know, if this is "urgent" anymore:
> My i686 is in fact a PII: perhaps someone may run "Haney Speed" built with
> today's 3.1 on an AMD core to exclude quickly the possibility of another nasty
> partial register stall???
Hm, up to my knowledge, the AMD Athlon didn't suffer (at least that much)
from partial register stalls (??) Athlon suffers from the higher latencies
of the VIA (or VIA alike AMD) chipsets (Why is VIA-KT266A faster
than KT266 ?) and his small TLB's.
Anyway, currently there is no gcc-3.1 installed on my system(s). Any
suggestions for a recommended version, that builds clean on a i686 (binutils
2.9.1.0.xx in /usr/bin/ & 2.11.2 in /usr/local/bin) and a i586-system
(binutils 2.9.1.0.xx only at the moment)?
Peter Schorsch
On Thu, 25 Oct 2001, Paolo Carlini wrote:
> Richard Kenner wrote:
>
> > Yes. I found a one-character typo in that change yesterday. I'm working
> > on
> > some other fixes before checking that change in. If I don't figure them
> > out soon, I'll just test with that change.
>
> Hi all again... bad news...
>
> Timing Real matrix operations for n = 200
> 1...2...
> Language Answer Check Time Iter Time
> O-O C++ (2.41602e+06): 28.9 0.289
> Hand-coded C (2.41602e+06): 14.84 0.1484
>
> Unfortunately, today, after
>
> Wed Oct 24 12:41:19 2001 Richard Kenner <kenner@vlsi1.ultra.nyu.edu>
>
> * emit-rtl.c (set_mem_attributes): Fix typo in last change.
> * print-rtl.c (print_rtx, case MEM): Improve display of MEM_DECL.
>
> that horrible slow down for C++ is still there...
>
> Suggestions?
>
> My i686 is in fact a PII: perhaps someone may run "Haney Speed" built with
> today's 3.1 on an AMD core to exclude quickly the possibility of another nasty
> partial register stall???
>
> Thanks,
> Paolo Carlini.
>
> P.S. My build switches for the Haney code were "-O2 -funroll-loops
> -fomit-frame-pointer -DNo_Fortran -D__KCC". My system is a PII-400, 256M,
> Linux2.4.10, glibc2.2.4, binutils2.11.2.