Mail Archives: pgcc/1998/05/13/17:29:08
Wolfgang Formann wrote:
>
> As I said, you already had 15% by using AMD (compared to 233-MMX-Pentium),
> so expect only the rest, giving a total of 30%.
if I may, I want to add something:
I have a 233-MMX-P5, a PPro-200 and a P6-233 (for me p6=Pentium II). If
you do the Byte-Benchmark (v2.1 I think ?) for all this machines, using
gcc from the Pentium group, you (well, at least me !) get floating index
(index=1 is based on a K6-233, using gcc 2.7.2.3) between 1.1 and 1.9
(the last one is for the p6), thus quite better that for the K6.
(and even with the same compiler and the best optimizations, I doubt
that you will get similar results).
NB: the results for the p6 was done with the snapshot pgcc-980406, and
the flags -mamdk6 -O9 -funroll-all-loops -fexpensive-optimizations.
The flag -mamdk6 for the p6 is not a joke ! I usually get the better
results with this flag, sometime it's 15%-20% faster !!! I don't know if
someone else also observed this behavior, and I don't know it this is
relevant for the "Pentium team". I also don't know why -O9 get a little
improvement on the result. Note also that the flag -pentiumpro is a
disaster for my p6 ...(Yes, I'm sure that there is a p6 in the
computer, and the motherboard is a QDI Legend V)
If I remember some Specf95 between K6 and Pentium, the index were always
better for Intel ...
Thus, depending of your applications, it's not clear that you will have
better performance with a K6, in particular FPU intensive applications.
hope this help, Ph.
--
Laboratoire Geometrie-Topologie et Algebre. UPRESA CNRS 5030.
E-mail: pev AT pev DOT math DOT univ-montp2 DOT fr | Phone: +33 (0)467143958
http://www.math.univ-montp2.fr/~pev | Fax: +33 (0)467143558
- Raw text -