OC'd 2.4C - Prime95 question

R

[email protected]

Weaksauce
Joined
Jul 16, 2003
Messages
79
P4 2.4C @ 3.25
Asus P4P800 Deluxe
1GB OCZ PC4200
Radeon 9800pro 128MB

I have the CPU voltage at 1.625. A 1:1 ratio on the CPU:RAM. I have messed with nothing else. Temps are fine. Idle around 40C, load at 50C. (there abouts). It runs all my games, and runs them brillirantly. 3DMark2k3 scores of JUST under 6000. Seems decent.

The question. I run Prime95 and it will run for about 2 hours, then I get an error and it says something along the lines of "rounding error, got .05 expected .04" Anyone know what causes this error and how I could fix it? More CPU voltage? Less CPU voltage?

Sorry if this is a common question. I searched first, but all I could find where threads telling people to search.
 
It means something can't handle the overclock. Prime95 just crunches numbers and lets you know if one doesn't crunch correctly. It's not a diagnostic tool.
 
I know it's not a diagnostic tool. If it told me what the problem was, I wouldn't be here asking you guys. What I'm looking for is if anyone has recieved this error before and knows what the likely cause of it is. Obviously SOMETHING is wrong with the overclock, but instead of spending hours tweaking each setting and rerunning the test, I thought I'd see if anyone could point me in the right direction.
 
try stressing your ram, and cpu individually. i use memtest86 to stress test my ram, and the small fft prime95 to test my cpu. i may have the same ram as you and mine is stable 1:1 up to 271fsb @ 2.8v. my first guess would be maybe your ram is unstable at that speed.
 
Calm down man! I agree, it is not a diagnostic tool, just a number crunching app that is telling you that at your current settings your processor is returning bad results on particular expression that is being evaulated.

Overclocking is not a matter of "here is your answer, do this and it will work". If you want to find the max stable OC it takes a lot time and effort, there is no "Holy Grail" answer. That being said, if it was my pc I would try increasing the voltage little by little, verifying that the cooling and power is adequate and clean. If those two things didn't help I would back the frequency down a little to see if the errors go away. Since overclockability varies so much even between chips in the same batch yours might not be capable of running error free at the speed you are trying to run it at.
 
[email protected] said:
I know it's not a diagnostic tool. If it told me what the problem was, I wouldn't be here asking you guys. What I'm looking for is if anyone has recieved this error before and knows what the likely cause of it is. Obviously SOMETHING is wrong with the overclock, but instead of spending hours tweaking each setting and rerunning the test, I thought I'd see if anyone could point me in the right direction.
Click to expand...

Yeah, I've gotten that error lots of times. It's saying, "Hey, I told your computer to calculate this and it came up with the wrong answer." Like I said, it's not capable of telling you what the problem is (specific error), it's just telling you something is wrong causing it to make a miscalculation.
 
1.625v is allready a little much for a 2.4 to hit 3.2. mine does 3.3 on defualt voltage.
 
Sorry if I sounded snappy, it wasn't my intention.

acascianelli:What is the normal voltage for the ram? is 2.8v giving it extra juice? I haven't tried that yet. Also, you mention that 1.625v is a little high for the 2.4? Hmm. I couldn't get it to post at 3.25 without raising the CPU voltage. Would increasing the RAM voltage help maybe? Let me ask this instead ... what is the highest voltage you think is safe to through at the OCZ ram? I've heard it can take a lot.
 
the support voltage for the ocz pc4200 i have is 2.8v, so i just set it to 2.8v and went as high as i could go. my motherboard will only go to 2.8v anyway. id try setting your vdimm voltage to 2.8v and lowering your vcore to 1.55 or 1.575v
 
You must log in or register to reply here.
Back
Top