Originally Posted by snorre
No, neither your benchmark numbers or the x86-secret review gives a realistic impression of the performance differences. I also suspect that you're running those benchmarks with all power saving features disabled, something that soyboy clearly didn't do even if it hampered the true performance of his system. A fair performance comparion shouldn't include any tweaking and other hidden tricks in my opinion.
um, i would want all systems running at max power settings for these benches. i would think you
would too, since evidently are you still trying to downplay the performance of the dothan? how often do you suppose you would employ your a64 3200+ in limited battery life situations, thereby employing powerstepping features, in comparison to a dothan system that gets 2-3 hrs more battery life? anyway, this may be difficult to grasp for you. let's stick to your desktop-system-to-notebook-system comparison.
and you're comparing the dothan mobile chips to desktop system configurations where uh, there's always max power situations (rincewind pointed this out earlier). isn't it only fair to give the dothan its max speed settings before setting out to bench it?
no, those other chips are not downclocked in the example that rince posted. you didn't read his post correctly, nor have you really looked at what those french guys did. here
is the page where they downlocked all the chips. you can see that the dothan beats them in many of the cases when they're clocked the same. you omitted all of them, cleverly.http://www.x86-secret.com/articles/c...n/dothan-5.htm
they downclocked the other chips to the same freq as the dothan to show inherent differences between the two architectures. if you can read french you'd know that, for example:il écrase littéralement l'architecture NetBurst à fréquence égale, ce qui est toutefois normal, cette dernière étant conçue pour les hautes fréquences. SuperPI étant basé sur un compilateur ancien, passons maintenant sous Kribibench qui exploite le jeu d'instruction SSE et voyons le résultat :
reads something like:it literally crushes NetBurst architecture at equal frequency, which is however normal, the latter being conceived for the high frequencies. SuperPI being based on an old compiler, let us pass now under Kribibench which exploits the instruction set SSE and see the result:
now you can move to an analysis of the highest clock freqs that all chips can produce (this involves OCing the dothan to 2.4):
amazing, the dothan can peform almost as well as the power hog P4EE !