Overclocking Results

Overclocking

In previous tests, the only way we could get any of our GeForce 256 DDR cards stable at 133MHz bus on our BX testbed was to reduce the AGP Aperture Size. Rather than do this for the overclocking article, we decided to stick it out and see how fast the system could run before the AGP card finally gave in. While 124MHz was stable for a while, after repeated testing the system became unstable. Fortunately, the BE6-II features 1MHz increments for the FSB so we slowly dialed the bus down until we achieved complete stability.

After repeated testing, we finally settled on 120MHz. Only at that bus speed could we achieve complete stability. While 720MHz is a nice overclock, we were hoping for a bit more. With this in mind, we quickly set up our 820 testbed, based on AOpen's AX6C motherboard.

820 Testing

While the AX6C has several bus speed options and even supports manipulation of the RDRAM multiplier, one critical feature missing is voltage manipulation. For instance, our 600E system ran flawlessly at 138MHz FSB with default voltage, resulting in a clock speed of 828MHz. Once the bus was raised the system unfortunately became unstable.

We couldn't complete the full suite of tests without constant lockups and crashes in Quake III, which suggests the CPU could run at higher clock speeds with a little more juice. These feelings were confirmed during testing with the VIA chipset.

Apollo Pro133A Results

Since the 600E we borrowed was Slot 1, (Thanks Billy!) we couldn't run tests with our FC-PGA MS-6309 motherboard. Instead we relied on the ASUS P3V4X. Since we knew 138MHz was stable we next tried 142MHz. We had to raise the voltage to 1.8V to achieve complete stability, but the system was solid at 852MHz.

We next upped the bus to 146MHz. While the system ran well for a good portion of the Quake 3 testing, the system eventually crashed. A quick check of the BIOS revealed the CPU was operating at 130 degrees Fahrenheit which suggests insufficient cooling was the culprit. All tests were run with the Vantec P3D-5030 heatsink, perhaps we would have had better results with a Swiftech MC2001 "Coppercooler".

So what was the final clock speed we achieved with our 600E? 864MHz. (6.0x144) Quite an upgrade from a 600MHz CPU, and as our numbers show, this VIA Apollo Pro133A system is an impressive performer!