This is just a quick update. I tried 1125 MHz again, this time using a different strategy to overcome the voltage quirk of the CUBX board. I booted up at 133 or 140 FSB, setting the core voltage to 1.95V (measurement is almost 2V). For some reason, higher voltages will not POST. After booting into Windows I would use CPUFSB to set the FSB to 150MHz. This method proved to be surprisingly stable. The only problems is, I need to lower the FSB with CPUFSB before shutting down the machine, otherwise the voltage quirk will prevent the machine from cold booting. This happened to me once and I had to clear the CMOS RAM to revive the computer.
There is a published trick to overcome the voltage quirk and perhaps to enable higher voltages. It involves shorting the VID0...VID3 pins of the CPU and can be found under the name "VID pin trick" or similar. However, I am quite happy with the performance at 1050 MHz and default voltage, so this is not an option for me.
Actually, I just wanted to show a few quick pictures on request. The other parts of my overclocking report you find here. Just for reference, the P3-1GHz was running at FSB 150MHz on an Asus CUBX. Memory is 2-2-2-8. The CPU is lapped and cooled by an Alpha PAL6035MUC. No WPCREDIT tweaks have been used. Video card is a Hercules GeForce2 with 32MB and 200/366 timing.
So much for now,
Martin
It is hard to keep up with the constant driver updates by Nvidia, but I thought I share the latest 3DMark2001 score...
| < Part V | BIOS modding guide > |