System has become wobbly after months of stability

Associate
Joined
2 Jun 2005
Posts
260
I have an Asus A8N-SLI Premium mobo with the 1009 BIOS, an X2 3800+ CPU, 2x1GB Mushkin XP4000 black RAM.

After buying this rig before Christmas I had a stable overclock of 2500MHz. HTT x3, CPU at 250MHz 1.5V, RAM at 1:1 3-4-3-8, 2.8V. It ran Prime95 stable on both cores for more than 48 hours when I was setting it up. Up until March it was rock steady with whatever I threw at it.

Recently however I've been plagued by an odd behaviour that I would describe as paralysis. At random times after a few days of running, not just under heavy load either, my system would start to become unresponsive. Processes would still run fine until I tried to interact with them. For example, I was watching a DVD on my secondary display using ZoomPlayer and I noticed my apps on the primary display had frozen. The DVD continued to play fine until the film finished, at which point I tried to close ZoomPlayer. It had frozen like everything else. I could still move the mouse pointer and if I had multiple apps visible some would look OK until I try clicking on them, at which point they freeze too.

CPU usage was at 0% for both cores and the mouse could still be moved but everything was frozen, no disk or network activity. I could even press CTRL-ALT-DEL and all apps would disappear like it usually does before the window pops up with the Task Manager, Logoff, Shutdown and other buttons. However that window would never appear [left it over an hour waiting!]

At first I thought it must be a low-level driver as the PC was not completely frozen, the CPU was running cool [front panel] so it can't have been a standard rogue process. I uninstalled any software that had installed a low-level driver, including Daemon Tools, Diskeeper, PGP Desktop, Soundcard [Hitec Mystique DDL 7.1]. Almost everything except the Firewall as I considered running without it too risky.

Most of this was done with my overclock disabled, the BIOS settings set back to the defaults because I wasn't sure if that might be the cause. The problems still keep arising so now I'm thinking the CPU is damaged. I can't imagine RAM damage would make it behave this way with BSODs all over the place.

Has anyone else experienced this weird kind of 'paralysis' on their PC? I'm at a loss to know what to do next, besides risk buying a new CPU.
 
This is becoming stranger. As I have never encountered any errors using Memtest86 v1.6, even today I ran it for one loop, I decided to try HCIMemtest under Windows XP. It detects hundreds of errors with my RAM, and the PC BSODs within about a minute of starting it.

So then I modded the BIOS settings to change the divider from 400MHz to 333MHz. At the same time I also used the /maxmem= boot.ini option to limit Windows to 512MB. So far no problems, HCI has tested 'All Unused RAM' 2100% [I presume it means 21 loops].

I need to try it again with the full RAM available to be sure, but all these BSODs running HCI now finger the RAM as the cause.
 
So changing the memory divider from 400MHz to 333MHz is enough to make it stable. The RAM is now running at a sloth-like 167MHz [rated to do 250MHz and worked at that speed for months!] but coped with 16 hours of VirtualDub encoding straight with not a hint of a problem.

I think I might RMA the RAM. How much chance is there of it being anything else? Do RAM controllers on X2 CPUs fail at all?
 
Back
Top Bottom