In the Lab: Gordon Mah Ung Mulls Overclocking

One clear lesson that came from all the overclocking I did for this month’s cover story is that bargain processors kick butt over their pricier brethren when it comes to increased clock output. The most impressive pure clock-speed gain I saw was with the Pentium E2160 chip. That sucker jumped from 1.8GHz to 3.37GHz—roughly an 85 percent overclock! Our gauge of success wasn’t just a post and screen grab, either. I used Prime95 to stress-test the processors.

Perhaps the most disappointing results were with AMD’s older 90nm Windsor cores. These make up the company’s fastest Athlon 64 X2 CPUs, but neither the new Black Edition Athlon 64 X2 6400+ nor the X2 6000+ coughed up many extra megahertz. I saw less than a 10 percent clock bump from either chip. And I’m not alone. I spoke to an OEM representative who said his company was also seeing only minor overclocking results from its Windsor core processors.

Even though AMD has decided to unlock its X2 6400+ as a reward to overclockers, it doesn’t seem that there’s really any point to it. AMD’s 65nm Brisbane CPUs performed a bit better, with stable 25 percent overclocks, but they paled against budget Intel dual cores, which all allowed for overclocks of more than 50 percent. Intel’s Core 2 Quad Q6700 was another impressive processor. Although we saw but a 30 percent clock increase, that’s not bad for a $500 CPU, especially when the pricier Core 2 Extreme QX6850 doesn’t go that much farther. Of course, your mileage may vary. We decided to stick with standard air-cooling for all of our tests; water or phase-change/hybrid solutions would likely yield even better results.

I’ve long believed that the bulk of overclockers aren’t pushing $1,000 CPUs as far as they can go. It’s really the guys and gals buying the $180 CPUs who are trying to get as much bang for the buck as possible.