Finally, we attempted to reach the magical clock speed of 4 GHz. This time, we reverted to the 333 MHz FSB and used a 12x multiplier, giving us a clock speed of exactly 4 GHz. In our first attempt, we were just barely able to glimpse the BIOS POST before the motherboard froze.

Since the board wouldn't even boot with these settings, we knew that we would have to increase the voltage further - a lot further. With a heavy heart we set the voltage to 1.4375 V in the BIOS, which amounts to an increase of 0.1875 V. At this high voltage, we were not only able to get the QX9650 to boot, but also to complete the Prim95 stability test.

Still, running a 45 nm processor at this voltage is definitely not a long-term solution, since there is a risk that the CPU is gradually damaged as a result of electromigration. There is a possibility that the conducting paths inside the CPU could literally be ablated and thus destroyed. The result is that one day, the computer suddenly shuts down and will never start up again with that CPU.

At 1.4375 V, we were able to push our Penryn sample to 4.0 GHz.

Nonetheless, we must say that we were more than a little surprised that the processor would run at 4.0 GHz at all, let alone in stable fashion. We should probably mention at this point that the processor only used air cooling (via a Zalman CNPS9700 LED) for all of the overclocking tests, which just makes the results even more impressive in our eyes.

All of our tests were run with a Zalman air cooling solution.