Navigate: Advice Beginners BIOS Guide CPUs Links Mainboards Memory Network Storage Video/Sound Cards Contact Forum SiteMap Sponsors WebNews Home	.	.
Prices: Mainboards ABIT ASUS Chaintech Shuttle Soyo Tyan CPU Intel P4 2.4C-800 P4 2.6C-800 P4 2.8C-800 P4 3.0-800 P4 3.2-800 AMD AthlonXP XP 1700+ XP 2000+ XP 2400+ XP 2500+ XP 2700+ XP 3000+ XP 3200+ Athlon64 Athlon64 3200+ Athlon64 FX-51 Opteron Opteron 240 Opteron 242 Opteron 244 Opteron 246 Memory Corsair Crucial Kingston Mushkin OCZ Search Prices:

There is too much controversy about Sysmark at the present time for us to include it into the selection of benchmarks we are running. Granted that within every single application, there are operations that favor one or the other platform, we still abstain from Sysmark2002 just as we did with Sysmark2001 that required a patch to run the Athlon processors at their full potential which in turn was a violation of the licensing agreement explicitly stated by BAPCo. All in all, who cares, there are enough applications and benchmarks out there that can be used to choke any reader by raw mass. Likewise, we will not use SiSoft Sandra in this review since, thus far, the sense or nonsense of the CPU benchmarks has eluded us. That does not mean that either Sysmark or SiSoft are not valid benchmarks, on the contrary, both benchmarks contain very valuable aspects, only for the purpose of this review, they are either redundant or out of place.

PCMark2002

One of the new platform / system benchmarks is PCMark2002. Overall, I am still not sure whether the benchmark really provides any information that goes beyond showing some numbers but for the sake of the argument, we'll show the data anyway

Increasing the memory speed from 166 MHz (DDR 333 or PC2700) to 200 MHz (DDR400 or PC3200) increased the CPU scores by less than 1% which is what we would expect and gave 7% increase in memory performance compared to an overall increase in raw bandwidth of 25% which shows that the memory benchmark goes beyond just measuring raw data stream bandwidth. Increasing the FSB and, by extension, the CPU clock rate scales in a linear fashion with the CPU scores.

Content Creation Winstone 2002

If there is any benchmark that can be used at this time to really hammer on CPU, memory and I/O system simultaneously and which still provides some sort of relevance for what any user might be interested in, it is probably Content Creation Winstone 2002, even though not even "The Flash" would be able to input all data used by the benchmark within a full week. Still, it is conceivable that at some point there will be prefabricated application packages that will dump huge work loads on any system / processor. With respect to office applications, this scenario borders on the ludicrous which is why we are still not using Business Winstone, however, in the realm of content creation and the increasing possibilities as e.g. real-time Photoshop filtering of incoming video streams, there may be some future applications that could demand the amount of psocessor and system power that we are looking at right here.

Content Creation depends too much on I/O bottlenecks and CPU performance to show much influence of memory speed on the overall scores that are, nonetheless, rather impressive.

Caligari TrueSpace 5.1, Vase Scene

Runtime in seconds, shorter is better. Thus far, the chapterhouse benchmark for AMD processors, we see the 2.8 GHz P4 getting some serious scores. For comparison we throw in scores for the P4 2.53 and the Athlon XP2100+ (Palomino) and XP2200+ (Thoroughbred). Needless to say that the overclocked P4 bags it all.

Even at clock speed, the 2.8 GHz P4 beats the fastest AMD CPU we ever tested hands down.

next page:    => Benchmark Results: Gaming Performance =>