prime95 error

FATAL ERROR: Rounding was 0.4990234375, expected less than 0.4
Hardware failure detected, consult stress.txt file.

I got this error on the 7th hour on one of my cores on blend test

Here are my settings:

AI Overclock tuner: Manual
CPU Ratio Setting: 9
FSB Strap to North Bridge: AUTO
FSB Frequency: 400
PCI-E Frequency: 100
DRAM Frequency: DDR2-801

DRAM Timing Control: AUTO
DRAM Static Read Control: AUTO
DRAM Read Training: AUTO
MEM. OC Charger: AUTO
Ai Clock Twister: AUTO
Ai Transaction Booster: AUTO

CPU Voltage: 1.3625 <------------(max safe voltage, im afraid to push this)
FSB Termination Voltage: 1.1
DRAM Voltage: 1.8
NB Voltage: 1.2
SBridge Voltage: 1.5

CPU GTL Reference: AUTO
CPU Spread Spectrum: Disabled
PCIE Spread Spectrum: Disabled
CPU Clock Skew : AUTO
NB Clock Skew : AUTO
Enhancement: Performance

c1e, eist, disabled

i have an asus p5q se2 motherboard, a C2De7500, and 2 DDR2-800 ram.
Really old pc ( well not old, but the specs are)
Tell me if i need to post more hardware for you


Any idea how to fix this problem, my buddy says its a voltage error

Odds are that your buddy is correct, leaving you with two basic options:
1) More voltage.
2) Lower clock speed.

However there are a couple things you can try, the first thing I would try is a bit more VTT (more common name for FSB Termination Voltage), try 1.2v rather than 1.1v.
That may do it right there if you're lucky. Another thing to try is using the 333 FSB Strap to NB, many P45 boards really don't appreciate using the 400 strap for some reason, especially Asus P5Q series boards. No clue why, but mine certainly didn't like it much.

If other things fail, try either dropping the FSB to 390 or raising the voltage to 1.3725ish.
Intel 45nm and 65nm chips are quite strong voltage wise, which way you go is of course up to you. Given that it made it six hours before having an issue, it is entirely possible that 399 or 398fsb would work, going that long implies that it is very close indeed to being stable. Another thing to look at is the actual voltage the CPU is getting, CPUz should tell you. If it is dropping a lot under load you'll need to turn on LLC (Load Line Calibration) or raise the vcore setting in the BIOS to compensate.

I think this board has no LLC. I cant find any under the BIOS

Would it be ok to try all at the same time( 1.2v on FSB, 333 FSB Strap to NB, and up the vcore to 1.3725)

Max vcore on cpu-z is 1.344. Ill try raising the vcore to 1.3725. I hope it wont get past 1.3625

Is 1.2 safe for Vtt?

Sure it's OK to try all. You might have to do all to get it to stabilize. You're close, really close. Being that you are so close, I'd change one and only one variable at a time.

You can definitely do it all at once if you want to. The only downside is not knowing which part worked if it is successful.
If you can live with that, go right ahead! It'd drive me nuts personally

Wooot! tried uping the core to 1.3725 and cpu-z max reading is 1.352

I cant really stress test for 7 hours in the morning. im too impatient. i think ill try all of them for stability. Since i use this comp for other purposes

Ill leave it on overnight for each settings.

oh yeah! 24 hours on blend test and no errors

now for memtest

I wouldn't bother with memtest. I'd consider it to be stable. Now to OC your GPU.

mmmm i think i dont need to since both my 560ti can run almost all at ultra

plus its only on AIR. dont wanna risk it room temp in our country is 30c

In that case get some beer and celebrate.

Cigarsmoker said:

In that case get some beer and celebrate.

Click to expand...

cheers

Please get Intel Burn Test and run that for about 10 minutes to see if that passes.

http://www.xtremesystems.org/forums...-IntelBurnTest-The-new-stress-testing-program


If this fails, you probably require another Vcore bump.