• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

195 memtest86 errors. Wtf?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

jEevion

Member
Joined
Oct 25, 2004
Location
Ontario, Canada
Hi, I have been having frequent application crashes lately, so I downloaded memtest86 and ran it. After 1 complete pass, it detected 195 errors. My specs are:

AMD64 3200+ @ 2.5GHz 10x250
OCZ Dual Channel Platinum Rev. 2 @ 250MHz 2.5-3-3-7 1T
OCZ PowerStream 520W
MSI K8N Neo2 Platinum

That should be enough info. If I loosened latencies, could I get rid of these errors? Or does this mean my memory sticks are completely bad and need a replacement? Thanks a lot.
 
Your memory may not be able to hold those high clocks. Are you running the 3200EL Rev. 2's?

Then if loosening the timings doesn't help. You probably need to lower your clocks a bit or else try running 2T, whichever ends up being faster overall.
 
I have the OCZ EL Dual Channel Platinum Rev. 2's
They are at 2.85v, now at 3-4-4-8 timings.

I thought that this memory should be able to go up to 260MHz with no problems? I'll try decreasing the HTT to 245/240 and run the test again.
 
I just ran them at 200MHz stock at 2-2-2-5 and got 61 errors on test 4. Does this mean that I have to RMA the memory?
 
You may want to try them on someone else's system to check whether they error there as well.

If it does error again, unfortunately you probably should RMA.
 
Though only if the 2-2-2-5 at 200Mhz are in fact specified by OCZ at their recommended voltage...

Errors in memtest are an ominous sign. Running a system that causes these errors will not be a satisfying experience--at least that's what I've logged.
 
Try running it a speed lower, ie if you have PC3200 ram run it at PC2700. Along with that, loosen your timings a bit. That should be a temporary stable solution, but if your getting errors on test #5 in memtest86 - you have faulty memory that can't run at it's advertised speed. Time to RMA.
 
Well, I tried each stick individually, and indeed one stick was getting errors. The other is fine, I have it in now and it's working like a charm. I am going to RMA them Tuesday, I can't send them off Monday, no couriers are working. It was close, I bought the memory Dec. 1, and you get 30 days to make an RMA request, made mine on the 31st lol. But one thing ....

After removing one stick, I noticed that whenever I reboot my system, it sometimes hangs there. Sometimes it hangs right away, others it hangs at the Windows loading screen. Is the following a possibility.....

When I removed my sticks, since I have an XP-90 and a vantec tornado 92mm case fan on it, it's very crowded and takes co-ordination to remove and replace memory modules, as they are right under the heatsink. When I was putting them back in, could I have pushed the board down so that the heatsink lost some of it's contact with the processor? Or something to that nature? Or is it my MSI board? Thanks a lot for the help.

EDIT: Also, my CPU temps used to be 34idle, now it's 38idle. I hope I didn't screw anything up ... could this be - since there is less ram, the processor makes more effort to handle processes, thus temps rise slightly? Thanks again. When I had 1024mb memory, CPU temp never went over 42, and I only have the fan at 1/3 speed.
 
Last edited:
The CPU will not get hotter by itself if you remove some ram. You may have dislodged the heatsink.
Still, since you are talking about the PC hanging at boot or later at times, it seems to me the system isn't quite stable yet. Did you test the RAM sticks in the same slot or did you leave them were they were? It may be that one of your mainboard slots is faulty instead.
 
Sorry to resurrect this thread, but I must say something.

Before, I was complaining about my CPU temps raising after I put in old memory. What I did was, left the VCore at 1.56v and clocked the processor back to 2.0GHz. Even after doing this, my IDLE CPU temperature was 40 degrees.

I had to RMA my OCZ Dual CHannel Platinum Rev. 2 .... I got my new sticks, they work awesome, and I overclocked back to my recent 2.6GHz. And guess what .... MY CPU TEMP IDLE DROPPED TO 35 DEGREES. So in summary:

When the CPU is at stock speeds with high voltage (1.56v), my idle temp was 40 degrees. However, when clocking to 2.6GHz, 1.56v, the temperatures DROP to 35 degrees idle. Now what the hell ..... that's odd.
 
Idle temperature readings aren't very informative. Slight variations in CPU load (perhaps you've installed programs that have 'agent' processes running) may cause higher temperatures. Try to obtain a CPU load baseline instead; by stressing the CPU in the same way for a measured interval you reduce the error.
Further sources of error are the measurement itself: did you measure it with a probe or did you rely on some software tool (software can't be trusted).
Did you take variation in room temperature into account?
 
Back