Barton Review @ Extremetech

February 19th 2003 | CPUs & Chipsets

ExtremeTech recently reviewed the new Barton core Athlon XP processors, giving details on the prices, the performance and the proficiency of the processor. Complete with comparisons to Pentium 4 and Thoroughbred B core Athlon XP processors, using the following benchmarks:

3D Rendering
Lightwave and 3D Studio

Media Encoding
WME9 beta, Sound Forge 6.0, Adobe After Effects, Moviemaker and Quicktime 6 Pro

Others
PC Mark 2002, 3D Mark 2001SE, and various gaming benchmarks

Rather loaded, I’d say. Click Read More for quotes, or the link at the bottom of this article for the review.

Barton Review @ Extremetech

PCMark’s CPU tests tend to be cache bound — but apparently fit into a 256KB L2 cache. This would certainly explain the Athlon XP 2800+’s better performance in this synthetic test, as it’s clocked higher — 2.25GHz — than the Athlon XP 3000+, at 2.16GHz. However, the overall memory score for the Athlon XP 3000+ is better. We’ll take a closer look at this result below.

The 3D Mark test, on the other hand, likes the presence of the bigger L2 cache — enough so that the Barton posts a significantly higher number than the older Thoroughbred. But neither Athlon is faster than the Intel 3GHz processors.

The situation reverses a bit with Unreal Tournament 2003. The Athlon XP 3000+ easily bests the Pentium 4, while the 2800+ lags. Note also the difference in results between the 7205 chipset and the 850E in this particular game, possibly because of significant branching in memory and reduced page access latency for DDR vs RDRAM.

Here, AMD is introducing a product that clocks slower than its previous top of the line, but with more cache. The net result is that applications that are sensitive to cache size may run better on Barton, but apps which like higher clock rates (and have tight loops running in L1) will run faster on the 2800+.

Barton Review @ Extremetech
Published in: CPUs & Chipsets on 2003-02-19