Testing the Processors
This content was originally featured on Amdmb.com and has been converted to PC Perspective's website. Some color changes and flaws may appear.My first test was to re-overclock the original 2200+ processor I used in my initial review. However, also by request of many individuals, I used the Gigabyte motherboard that AMD supplied with the Thoroughbred processor. There was some thought that perhaps the Epox 8K3A+ motherboard had some features or power benefits that other reviewers might not have had. Even with the new motherboard, I saw nearly identical results, with the frequency reaching a staggering 2123.64 MHz, just below is what I was able to pull off on the Gigabyte KT333 motherboard.
This next image shows a rather lengthy run of the SiSoft Sandra Burn-In Wizard running on the same machine as the 2200+ processor running at 2123 MHz! The system had been running for over 6 hours at this point!
Overclocking the 2100+ Thoroughbred processors also turned out to be a very interesting endeavor. The first one, which we will refer to from this point on as 2100+A, is of the same core design, process and from the same fab as the 2200+ we were working with. Not much to my surprise, the fastest I was able to push the processor was 1.85 GHz: only a 117 MHz overclock. Compared to the results of the Athlon XP 2200+ we used, this is low scaling from the processor, but does fit closer to what other hardware editors have seen with their product. Here is the WCPUID screen from the 2100+A processor:
Our third and final processor to be tested, called 2100+B from now on, showed to have better results than 2100+A. This CPU was overclocked to 151 MHz FSB with a 13.0x multiplier to get a total frequency of 1.967 GHz, just over a 200 MHz increase in speed. These results are interesting to see as the fact that the two 2100+ processors range is so broad. Even with their near identical fabrication (according to the numbers we see on the CPUs themselves), they differ in overclocking potential by 75% or so. This makes it very difficult to make any kind of broad judgment on the processors scalability.