Pentium4 vs. AthlonXP, Clock for Clock - Benchmarks
(Page 2 of 3 )
Pentium4 vs. AthlonXP, Clock for Clock Comparison
BENCHMARKS (continued):
SiSoft Sandra Memory Bench:

Now here we actually see the opposite of what we've seen so far, the P4 is eating the AMD for lunch in memory bandwidth. The logical deduction here is that the Intel 845PE chipset is just basically better at handling the memory requests that the Via KT400 chipset on our AMD system, but this warrants mentioning. The really odd thing here is that when I bench the CPU at default (12.5 x 166) it is right in line with where PC2700 memeory should be. Rather the chipset has a problem "underclocking" the front side bus of the CPU or not I cannot tell you.
PiFast Bench:
This benchmark actually requires a little bit of explanation. I consider this to be the most telling benchmark we are showing you today. Pifast measures basically the CPU's raw performance and nothing else. Pifast does exactly what it says it does, it calculates Pi. Obviously, the lower the time the better the score. We used the options as laid out below:
Program : PiFast version 4.2, by Xavier Gourdon
Computation of 10000000 digits of Pi
Method used : Chudnovsky
Size of FFT : 1024 K
Physical memory used : ~ 61355 K
Disk memory used : ~ 0.00 Meg
Here are the results:

Wow, that's a pretty significant improvement from the Pentium4 to the AMD. I actually ran this three times to be certain of the benchmarks accuracy, all three were in this neighborhood. With the SiSoft benchmarks the naysayer can blow off the results as a synthetic benchmark. With PiFast we are asking the CPU's to perform calculations and the AMD simply blows the P4 out of the water, there's no other way to see it.
QuakeIII:
Now here was a curious choice you might think for a 3D benchmark of the CPU's at like speeds. Why QuakeIII and not a newer, more stressful game you might ask? For two reasons really. First, QuakeIII remains a staple in the PC enthusiasts community as a benchmark tool of choice. Second, it is an older game that quite frankly can't put much smack down on the video card, ruling the video card out as the bottleneck. We wanted to see the performance of the CPU, not the video card, so QuakeIII was our choice.

Now here is something that is rather interesting. After watching the AMD easily handle the P4 in prior benchmarks, they break dead even here. Whether that's because both of these systems are pushing Q3 as hard as it can go, or that they are truly even on this mark I can't tell you, but there it is all the same. I can tell you that only a 30 FPS difference between 1600x1200 and 640x480 is rather impressive for both. :)
Let's head to the final page and check out 3dMark2001 scores and wrap this thing up...
Next: Conclusions >>
More Computer Processors Articles
More By Jim Miller
| Recommended by Dev Hardware |
|---|
|