Overclocking

All kinds of P3-600

As we mentioned before, there are two different Coppermine P3-600 processors, the 600E and 600EB. The only differences between the two processors are the different FSB speeds, but the FSB speed plays a huge role in the overclockability of the processor. Remember that Intel clock locks the multiplier for each processor, and the only way to overclock an Intel processor is to increase the FSB speed. Increasing the FSB speed affects more than just the processor. Tweaking the FSB stresses the entire system bus, everything from the system memory to the PCI and AGP bus.

The 600E's 100MHz bus limits the chip to a 6X multiplier (6 x 100MHz = 600MHz), and the 600EB's 133MHz bus speed limits the chip to a 4.5X multiplier (4.5 x 133MHz = 600MHz). That makes the 600E the best choice for overclockers because the relatively low 100MHz bus speed leaves a good bit of room for overclockers to increase the bus speed.

Overclocking the 600EB will be more difficult because, in order to increase the already hearty 133MHz bus speed, you need system memory that can handle bus speeds higher than 133MHz, and you'll also need an 1/2 AGP divider otherwise your AGP video card won't be able to handle the increased AGP bus speed. BX motherboards only have the 1/1 and 2/3 AGP dividers. Of course, the BX chipset was never meant to officially support the 133MHz bus speed.

Engineering Samples

We were fortunate enough to get our hands on three P3-600 FC-PGA engineering samples. As many of you already know, Intel's engineering samples are not clock-locked, which means that they'll work with any clock multiplier. We used an Abit BE6-II motherboard in our overclocking tests because the BE6-II's Soft Menu III allows users to adjust the FSB speed in 1MHz increments from 83MHz to 200MHz.

After testing all three processors, we found that one was stable at 750MHz (7.5 x 100MHz), and the other two were stable at 700MHz (7 x 100MHz). The 700MHz processors simply couldn't POST at higher speeds, and attempts at 750MHz allowed us to become very familiar with the motherboard's CMOS reset jumper. Adjusting the voltages didn't help, and we decided to run a couple other tests.

We decided to test the 750MHz CPU with the 4.5 multiplier to simulate a 600EB. The system couldn't handle the 166MHz FSB speed necessary for 750MHz (4.5 x 166MHz). Our PC-100 Hitachi SDRAM that we've had running as high as 150MHz couldn't handle 166MHz. Just to make sure the system failure wasn't the fault of the AGP card, we yanked out our GeForce and replaced it with a PCI TNT2 card (ASUS PCI-3800), and the results were the same.