Ram causes graphical glitching...?

[siberx]

I've been having a strange problem trying to o/c the memory on my system. Firstly, this ram was previously used in an nforce2 based system, and in that case worked stably in dual channel past 225mhz (as fast as the board could handle) at 2.5-3-3-11 and 2.6v. However, when using this same ram in my new system, if I clock the memory bus up beyond about 215, graphical corruption shows up on the screen whenever I drag windows around (see attachment). This happens with increasing severity up to about 235fsb, beyond which point I am unable to stably boot. If the system is exhibiting this graphical corruption on the desktop, it will also blackscreen after running a 3d app for a few minutes.

 

I have tried just about everything I can think of in the bios, including playing with all the memory timings, HTT bus settings, PCI-E settings, and just about anything else that could have any relation to the memory or graphics at all. Although the primary timings (2.5-3-3-7) seem to have no effect on the corruption, disabling DRAM Bank Interleave reduces the errors somewhat, but does not eliminate them completely.

 

I have tested the machine to be prime95 stable at 230fsb, even when the corruption is occuring. A memory divider will alleviate the problem, so long as the divider puts the memory below about 215fsb (i've tried HTT buses up to 280 to be stable, so it seems only the memory bus is affected). Also, I was able to reproduce this in another system with my board/proc and mosel vitelic memory (although corsair instead of OCZ), and an x800xl graphics card.

 

Has anybody seen or heard of an nforce4 based system that is completely stable, prime95 or OCCT-wise, but causes such severe graphical corruption? Any ideas on how I could possibly eliminate this corruption to allow me to overclock my memory properly? This is driving me up the wall!

[calvintang]

does your system freeze?if it really bothers you,you might wanna try to rma it

[siberx]

Well, it's not a problem with my hardware specifically, as I was able to cause the same thing to happen on another very similar system (same board/proc and same ram chips) - and it's only causing these problems while overclocking (the ram is perfectly fine at 200). The thing that bugs me about it is that the graphics go all funky from the memory bus being too high, yet the system is still fully prime95 stable. So what's causing the problem if it isn't the ram, according to stability tests?