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:

LOSTCIRCUITS	SHORTCUTS: New Performance Dimensions Test Configuration Cachemem Sandra Everest Analysis-1 FarCry The Poodle's Core More Benchmarks Conclusion Give Us Some Feedback to Help Us Improve our Reviews
Intel Pentium4 Prescott / 925XE chipset New Dimensions in Performance
(Review by MS November 21, 2004)

FarCry

We ran several benchmarks that returned more or less the same results, mirroring the memory performance data. Since rumor has it that in some games the internal frame counter is skewed, we used FRAPS to support our claims. Unfortunately, though, it turned out that the runtime of most benchmarks was too short to generate reasonable FRAPS logs, we'll explain the reasons for that in a bit. In addition, our goal was to run consecutive loops of the same benchmark without interruption and one of the benchmarks that combined exactly this feature with a resonably extended runtime is our own recorded LC5-Steam benchmark for FarCry.

Plot of the FRAPS values of three consecutive runs of LC5-Steam at 1024 x 768 x 32. As expected, after the first run is finished, the consecutive runs are overall faster. If you look carefully, though, you can see that towards the end, the third run gains additional speed. The three graphs were obtained after applying a "fix" to the system (we'll explain later)

We show the above plot to give an idea of how the system is supposed to run under normal circumstances. In the original configuration, however, the system showed the following FRAPS results:

For those staring at the graph in disbelief, yes, the numbers are accurate, the system that was running before at about 90-110 fps is now cruising along at 350-700 fps - according to FRAPS. The reason that some of the peak values are not shown is that the benchmark now only takes 13 (first run) to 10 (third run) as opposed to 50 seconds. Since FRAPS only samples once / second after counting all frames generated within that second, the graph suffers from a severe lack of granularity but this was the best we could actually do.

The same issues were encountered with Comanche4, MDK2 and a number of other benchmarks that were running somewhere between 4 and 8 times faster than on e.g. the same system with an Extreme Edition CPU. What is interesting, though, is that we could not see the differences when we looked at the benchmarks, nor was there a real difference in runtime.

next page: => The Poodle's Core =>