Yeah I was suspecting u were on 64 bit. Maybe the regex engine just isn't
optimized for 64 bit qwords. If it's just running 32 bit words then 50% of
ur XOR capacity is wasted. But then u would think that u'ld beat me on pure
clock speed. I seriously doubt that a Celeron is "better" in any way than a
normal Pentium. :P However, this Celeron has the Tualatin (sp) core rather
than the older one.
13*8 = 3.25 DWORDS
13*8 = 1.625 QWORDS
I did another very interesting test. I have both ActiveState and Cygwin
compiled perl on here. The Cygwin one is WAY faster at substr.
This is perl, v5.8.6 built for cygwin-thread-multi-64int
Rate RE substr
RE 1371742/s -- -53%
substr 2918856/s 113% --
This is perl, v5.8.6 built for MSWin32-x86-multi-thread
(with 3 registered patches, see perl -V for more detail)
Rate RE substr
RE 1302083/s -- -33%
substr 1937984/s 49% --
Now the second one is the exact same perl and the exact same code as the
test from my last email, but now the substr has mysteriously sped up. :D
At 08:43 AM 3/3/05 -0800, $Bill Luebkert wrote:
>That's interesting that your RE runs faster than mine and your CPU is
>only 1.4.
>
>I'm running B811 on an AMD Athlon 64 3200+
>
>Removing the $x++, I get (for longest string) :
>
> RE 949668/s -- -72%
> substr 3441156/s 262% --
>
>May have something to do with my 64-bit architecture or the fact
>that your on a Celeron rather than a normal Pentium.
>
--
REMEMBER THE WORLD TRADE CENTER ---=< WTC 911 >=--
"...ne cede males"
00000100
_______________________________________________
Perl-Win32-Users mailing list
[email protected]
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs
