Barton's Here: Athlon XP 3000+ vs. P4 3.06 GHz

Benchmarks Under Windows XP

We performed several different benchmark tests in order to obtain the most complete, well-balanced view of how the AMD XP 2500+, 2800+ and 3000+ perform, and of the overclocking results (2250 MHz, 2333 MHz and 2500 MHz). We still don't know what model number AMD will give these CPUs. The AMD roadmap clearly shows, though, that the Athlon XP 3200+ (based on the Barton core) will be on the shelves at some point. The benchmark results provide a clear overview of a total of 20 different CPUs. All new AMD Athlon XP processors have been included for comparison purposes.

You will also find that many things have changed in our benchmark tests. OpenGL performance was measured by different Quake 3 tests; Direct3D performance from the DirectX package was measured by the 3D Mark 2001 SE (based on DirectX 8). Nor did we forget Unreal Tournament 2003 and Comanche 4, both established 3D games for DirectX8.

The different MPEG-encoding benchmarks provide a comprehensive testing environment - "mp3 Makers Platinium" was used to encode a 178 MB WAV file into "MPEG-1 Layer 3 format." In a new discipline, we converted a DV video (1.2 GB) into an MPEG-2 file using Main Concept 1.3. We also created an MPEG-2 film using the latest version of the entry-level video-editing software "Pinnacle Studio 8.3.18."

A regular in our list of benchmarks is determining rendering performance using Newtek's Lightwave 7.5, 3D Studio Max 5.1 and Cinema 4D XL 8.001. The last two software packages have been updated in the past two weeks, allowing us to integrate the latest versions.

We also ran the new Winrar 3.11 archiving software to test how well the CPU performs when archiving files, a common application in the computing world. The Sysmark 2002 benchmark was used to determine office performance.

SiSoft Sandra 2003 is another standard in our benchmark repertoire. PC Mark 2002 is of interest to overclockers in particular, for finding comparison values for overclocking performance.