Quote Originally Posted by minsc_tdp View Post
Well I think I've screwed up my RAM. The system is now Orthos Blend unstable (bombs immediately) after approximately 8-12 hours of idle operation. I come home and my VMWare is spewing errors and the system is generally wonky. The VMWare instance at 128 MB RAM is new so it could be an issue but the problem seems hardware related.

I've gradually reduced my volts and clocks to near stock. All stock on the CPU, the mem is around 2.2 or 2.3V. Everything in my gut says the RAM is unstable. It's not the CPU clocks so now I'll need to start over and gradually work them back up again. The problem is somewhat more severe with high CPU clocks (380+) but it doesn't really have much of an effect vs. 266 MHz. So it's clearly the RAM.

The G.skill site says 2.2-2.4V. When I started it was OK at 1.8V and I quickly got it up to 2.2V. I eventually got to 2.4V and it's not helping this problem. How much further should I push it? I feel like it's a wild goose chase though. I also noticed after pushing the ram to 2.4V that a quick jump down to 2.2V was not bootable, it had to rest for a while. It seems most comfortable around 2.38V but right now it's just never stable.

I think I should start the RAM at the current 2.38V and beat on it with memtest for a minimum of 18 hours and try increasing the volts again if it is unstable, maybe going past 2.4 even though it's not even overclocked. I'd like to find a quicker way to determine that the memory is bad (maybe OCCT?), it's very odd that the system has to sit idle for 8 hours before the problem kicks in.
Sorry to repeat this but I really need some advice, it's driving me mad...