Prime fails small FFT after 1 hour 19 minutes.

[markp1989]

when i first set up this oc i only did an hour of blend, i was gona get ranked in other forum, but they wanted 2 hours small FFT, where now i fail at 1 hour 19min

specs are the desktop in my sig.

Vcore: 1.44 at load (cpuz)
fsb: 320
multi 12

Il go on my laptop and post the rest of the settings to see if you can help do somting to get it 100% stable.

 

[redduc900]

Fill out this BIOS template for the P45-T2RS w/ all of your current settings...

CPU Feature:
Thermal Management Control: Disabled
PPM (EIST) Mode: Disabled
Limit CPUID MaxVal: Disabled
C1E Function: Disabled
Execute Disable Bit: Disabled
Virtualization Technology: Disabled
Core Multi-Processing: Enabled

DRAM Timing:
CAS Latency Time (tCL) 5
RAS# to CAS# Delay (tRCD) 5
RAS# Precharge (tRP) 5
Precharge Delay (tRAS) 15
All Precharge to Act AUTO
REF to ACT Delay (tRFC) AUTO
Write to PRE Delay (tWR) AUTO
Rank Write to Read (tWTR) AUTO
ACT to ACT Delay (tRRD) AUTO
Read CAS# Precharge (tRTP) AUTO
Read CAS# Precharge (tRTP) AUTO
All PRE to Refresh AUTO

Voltage Setting: 
CPU VID Special Add: +87.5mV
DRAM Voltage Control: 2.194
SB Core/CPU PLL Voltage: 1.55v
NB Core Voltage: 1.10v
Clockgen Voltage Control: 3.45v
CPU GTL 0/2 REF Volt: 0.67X
CPU GTL 1/3 REF Volt: 0.67X
North Bridge GTL REF Volt: 0.67X
FSB Vref: AUTO

AC Shutdown Free: Enabled
O.C. Fail Retry Counter: 1
O.C. Fail CMOS Reload: Disabled
CPU Clock Ratio: 10X
CPU N/2 Ratio: Disabled
Target CPU Clock: 3600MHz
CPU Clock: 360 MHz
DRAM Speed: 333/667
Target DRAM Speed: DDR2 721
PCIE Clock: 100MHz

CPU Spread Spectrum: Disabled
PCIE Spread Spectrum: Disabled

 

[markp1989]

thanks for reply , here are all my current settings

CPU feature-------------
Therman management control [enabled]
PPM(EIST) mode [disabled]
Limit CPUID [disabled]
C1E function [disabled]
Execute disable bit [enabled]
Core Multi-Processing [enabled]

Dram timging ------------
every thing is set to auto

Voltages -----------------
CPU VID special Add [+337.5mv]
Dram Voltage control [1.908]
SB core/ CPU ppl voltage [1.55]
NB core voltage [1.195]
CPU VTT voltage [1.25]
Vcore droop control [enabled]
Clockgen voltage control [3.45v]
CPU GTL 0/2 REF volt [0.67X]
CPU GTL 1/3 REF Volt [0.67x]
North Bridge GTL REF volt [0.67x]
FSB Vref [Auto]

Genie Bios settings
Ac shutdown free [enabled]
O.C. Fail retry counter [1]
O.C. Fail CMOs reload [Bank 2]
CPU clock ratio [12.5X]
CPU N/2 Ratio [enabled]
Target CPU Clock 4000 mhz
CPU clock [320mhz]
dram speed [266/667]
Target Dram speed DDR2 802
PCIE Clock [100]

CPU Spread Spectrum [Disabled]
PCIE Spread Spectrum [Disabled]

 

[EarthDog]

Small FFT is usually CPU. Try giving is more Vcore and more Vtt.

 

[redduc900]

Try changing Vcore Droop Control to Disabled, both CPU GTL 0/2 REF Volt and CPU GTL 1/3 REF Volt to [0.63x or 0.61x] (0.67x should be fine, but you may still want to try 0.63x or lower), w/ North Bridge GTL REF Volt to either [0.63x] or [0.61x] (possibly even 0.58x). Although at such a low FSB of only 320 MHz, none of the GTL's should really need to tweaked, but they're also NB strap dependent. You might also want to play around w/ the clock skews, for example... CPU Clock Amplitude: 800mV, CPU Clock0 Skew: 200ps, and CPU Clock1 Skew: 0ps.

 

[EarthDog]

Whoa, changing the wheel already at such a low FSB? That thing natively supports 333Mhz FSB and he isnt even there yet...

 

[redduc900]

Just throwing some ideas his way.

 

[EarthDog]

And they are good ideas, but more a much higher FSB.

Actually after reading the thread I am confused. The title says Small FFT and the first post says blend...which is it?

 

[markp1989]

And they are good ideas, but more a much higher FSB.

Actually after reading the thread I am confused. The title says Small FFT and the first post says blend...which is it?

its small fft that fails, i fixed the typo

i have tried bumping the vtt and the vcore 1 notch up (cpuz still says the vcore is 1.440 at load :S) , and im testing, so far its been 20 min, il post back in a few hours how it all goes.

edit: i was under the imprestion that if an oc isnt stable that prime would error out alot earlier then an hour?

 

[Neuromancer]

Prime is actually slow to detect instability, which is why people normally say 12/24 hour prime is stable.

Other programs can detect instability in minutes that pass hours of prime.

OCCT, IBT, LinPack (which is an option in OCCT now)

 

[markp1989]

this time it failed after 1hour 21 minutes

edit: i dont get how it can happily fold for the last 5 weeks without issues, but not pass prime :S

 

[RJARRRPCGP]

edit: i was under the imprestion that if an oc isnt stable that prime would error out alot earlier then an hour?

Prime95 can take 2 hours to fail, sometimes 5 hours, AFAIK.

And Prime95 failing after 5 hours is RARE! That's what I call an oddball Prime95 failure.

 

[RJARRRPCGP]

this time it failed after 1hour 21 minutes

edit: i dont get how it can happily fold for the last 5 weeks without issues, but not pass prime :S

Prime95 is TOUGH! Prime95 is like the military!

Prime95 is the processor joining the military!

Prime95 can fail within 2 hours on a PC that looks perfect!

 

[EarthDog]

And Prime95 failing after 5 hours is RARE! That's what I call an oddball Prime95 failure.

False.

Prime95 is TOUGH! Prime95 is like the military!

Prime95 is the processor joining the military!

Prime95 can fail within 2 hours on a PC that looks perfect!

So what do you call IBT/Linpack/Linx if P95 is so tough?

 

[RJARRRPCGP]

I know that Prime95, with earlier versions, had a priority setting and Prime95 may be acting different for you, because of the priority setting not being set to "10".

For me, Prime95 was good at finding a problem within 5 hours.

 

[EarthDog]

I have the newest version.

A lot of issues do occur within 5 hours, but RARE I wouldnt say.

 

[freeagent]

Ive had prime go out at 8 hours :O

I rarely use it anymore, there are other apps that are much quicker at finding problems.. Prime is ok for helping your AS5 cure tho

I do like IBT

My cpu's dont

 

[Amtrak]

I've had Prime fail as much as 20 hours in. Prime's alright, but I also use lots of LinX. LinX is generally much better for finding voltages and tends to be much quicker at making an overclock fail. But just because it passes one doesn't mean it'll pass the other.

 

[RJARRRPCGP]

I've had Prime fail as much as 20 hours in.

LOL, like when Prime95 failed on me at 23 hours, because of the sun shining on my air intake. (with my Barton 3000+ AQZFA and Asus A7N8X-X) (duct from outside mod)

Prime95 would be OK all night, because the sun wouldn't ruin the show by making the temp go to 160 F or around that.

And once, I had Prime95 just disappear before, no error message and apparently terminated by Windows, because of overheating.

 

[Amtrak]

Can't say mine was from overheating, especially when it makes it through the hottest portions with ease.

I have, on several occasions with this case, accidentally turned the PC off more than halfway through a 24-36 hour stability test...
That's always aggravating.