See, this is the problem, and it’s fairly esoteric. I have a prebuilt Asus system that includes 9GB of memory on 3x2GB and 3x1GB sticks. The motherboard’s a Asus Rampage II Gene (LGA1336 socket or something like that) running an i7 930 CPU. I always, ALWAYS get freezes when running certain high memory demand applications. Crysis Wars is the latest culprit of this hard freeze (no error message, hard reboot needed).
This Christmas I got a nVidia GTX 460 1GB from my wife, and the installation there went smoothly. I had freezes with the GTX 260 I used to have in the system, but I thought it was driver related… Well, old drivers out, old card out, new card in, new drivers in, freezes now occurring more frequently in Crysis Wars. WoW runs fine, Windows runs fine. Crysis freezes…
So, I monitor my temperatures, I check the gauges and fan speeds to make sure I’m not cooking any of my components. Nothing. All I get is that the trigger point is somewhere around 45% memory usage.
So, I make a Memtest86 boot CD. It freezes at 4 seconds during the first test. The computer’s out of warranty, so any hardware issues will cost me money I won’t have unless I return the video card… To put it another way, Sad Fucking Panda.
I move on with the usual protocol: Remove and test all sticks individually. All sticks pass Memtest86. One momentary freak-out as the motherboard might be at fault, I test each individual memory slot. All check out fine on their own.
But… anything past 2 DIMMs at a time freezes Memtest86 at the 4 second mark. This motherboard has 3 memory modes: Single, double and triple memory channels. Single works fine, double works fine, triple freezes…
I’ve tried tweaking the memory timings, not the voltage. I’m not sure where to go from here. Reflash the BIOS? Find another setting to adjust? Crank the voltage up? Dump the new GPU? My googlefu has yielded no answer.
Yeah, like I said: