Prime95 Crashes computer

Toggle sidebar Toggle sidebar
B

beta124

Distinguished
Feb 2, 2008
59
0
18,630
So I'm trying to OC my Q6600 and things go great with memtest. I can run it for days.
When I try Prime95 (the x64 version), I can run it for several hours but then suddenly the computer will blue screen for a second and instantly reboot. Prime95 doesn't show any errors, it just instantly reboots the entire system.

Have you ever heard of Prime95 acting this way?
 
There are a few possibilities. It could be your motherboard FSB doesn't clock as high as you are trying to clock it. You might have inadequate voltages set. You could have bad memory......etc. What I would first try is to see if you memory is being funny using memtest86 and creating a bootable cd-rom and testing. Next I would say that you need to see what your MB FSB max is. I know gigabytes support 1600 mhz fsb by overclock. However, Abit won't support it on their IP35 PRO. You basically have to take your memory speed an multiply it by 4. However, first you have to half your memory speed then multi by 4. so for ddr2 800 you have to half it = 400 then multiply by 4 then you get 1600 mhz. The stock 333 clock speed of your processor x 4 = 1333 FSB. There are many things that could be creating the error.
 
I just thought that Prime95 would report an error on that core and then stop working that core. I didn't think it would reboot my entire computer.

Here's my system:
Cosmos 1000 Case
P5E WS PRO (BIOS V202)
Q6600 (G0)
Zalman 9700NT
XFX 8800GT
Corsair 8G 4x2G DDR2 800Mhz
Corsair HX750w
4x 500G Seagate HDD
XP Pro x64

Here's what she's running at currently:
Q6600 (GO) @ 3.01Ghz 333X9
1.2750v in BIOS CPU-Z shows 1.232v @ idle and 1.200v @ 100% load.
43-42-39-47 idle and 65-63-63-67 load

VID is 1.2875v

Memory is DD2 800Mhz running at 667Mhz 4-4-4-12 @2.1v

Ratio 1:1

So what do you think? By passing memtest for hours on end, I think my memory should be OK right? So what do you recommend for this processor? I'd like to keep the OC around 3.0Ghz.
 
When I think of it, I've had Prime95 quit a number of times, but it never caused the computer to blue screen and crash out, followed by an instant reboot. This sounds to me more like an overheat has occurred and the computer is resetting itself to prevent damage. Have you monitored the temperatures while running Prime95? Its something I always do, as I like to know how warm things are getting before things break down.
 
I always monitor the temps. Nothing looks out of place.
Here's a HWmonitor readout:


Voltage sensor 0 1.18 Volts [0x94] (CPU VCORE)
Voltage sensor 1 11.88 Volts [0xD5] (+12V)
Voltage sensor 2 3.30 Volts [0xCE] (AVCC)
Voltage sensor 3 3.30 Volts [0xCE] (+3.3V)
Voltage sensor 4 0.10 Volts [0xD] (VIN1)
Voltage sensor 5 5.06 Volts [0xD3] (+5V)
Voltage sensor 6 1.46 Volts [0xB6] (VIN3)
Temperature sensor 0 40°C (104°F) [0x28] (SYSTIN)
Temperature sensor 1 50°C (122°F) [0x64] (CPUTIN)
Temperature sensor 2 125°C (256°F) [0xF9] (AUXTIN)
Fan sensor 0 1140 RPM [0x25] (SYSFANIN)
Fan sensor 4 917 RPM [0x2E] (AUXFANIN1)

Intel Core 2 Quad Q6600 hardware monitor

Temperature sensor 0 61°C (141°F) [0x27] (core #0)
Temperature sensor 1 61°C (141°F) [0x27] (core #1)
Temperature sensor 2 60°C (140°F) [0x28] (core #2)
Temperature sensor 3 65°C (149°F) [0x23] (core #3)

GeForce 8800 GT hardware monitor

Temperature sensor 0 59°C (138°F) [0x3B] (GPU Core)

ST3250310NS hardware monitor

Temperature sensor 0 34°C (93°F) [0x22] (HDD)
 
I OCed my E2180 and yes if ur about right you get a red box and an error, but when i was off with voltages I got an instant reboot. Just raise voltages and watch temps. Also 2.75v 😱 .....I hope thats wrong lol.

Also do the temp Calibration stickied on the top of the section.....big help and knowledgeable. I was off by 5°C too low

Basically what iluv. just said below
 
look at it like this
Red box says fail means you are close to success OC(minor tweaks needed)
reboot or BSOD means you are very far off from success OC(LOADS OF TWEAKS NEEDED)
 
Well we need the VID from the bios not from a program like cpu-z because those are inaccurate. I know that the proc is like 1.3250 stock VID. I usually give it a slight bump in voltage. You don't have a problem with FSB cause all board support that FSB and your does too. 1333 FSB. Processor is stock at 1066 FSB. Anyways, I say it is voltage problem if those temps are under prime95 load. I would say that you should get the new coretemp

www.alcpu.com

Tell me what the temps are under load. Like I said. I'll bet it is a voltage problem.
 
As Gahleon said, bump up the CPU voltage, probably a minimum of 1.35v, or possibly up to 1.4v. If you haven't done it already, you might try bumping the ram voltage up to as much as 2.1v.
 
Well, I check in the BIOS and I can't find any VID readings.
But again CoreTemp (Version .99) says the VID is 1.2875v.

I think the max voltage for a Q6600 (GO) is 1.5v.
Do you think it's safe to go higher?
Here are my temps and volts under 100% Prime95 load:
BIOS 1.275v Idle
CPU-Z 1.232v Idle
CPU-Z 1.200V Load

43-42-40-47 Load
65-63-63-67 Load
By the way my processor is lapped and I'm using a Zalman 9700NT.

It will run for about 2.5hrs like that until the whole system reboots.
I can run memtest forever and no errors.

Memory is at 2.1v.
 
well from the temp you gave i would recommend you go any higher in fact you should lower it.or you lapping is not good enough that give such high temp.

i have OC a Q6600 before at 425x8@1.5v but in actual windows its lower.cant remember what it was.but as i said as long as you are in windows and idle vcore doesnt go above 1.5v in CPU-Z is fine.

and 70C being the max temp limit i would set for myself.some people say 65C too conservative i think and 75C is too much enthusiasm.

and prime 24 hrs max temp for is about 69-69-68-71.71C just for one core is acceptable in my view.

end of the day your call
 


With temps that high, you either need to back off the overclock or get a better heatsink. There are a number of tower types that will do better than the Zalman 9700. Check out some of the articles at Anandtech on heatsinks for guidance.

For a comparison, I've been running Prime95 for the past few hours and running a few business apps at the same time and my computer's core temps are all running between 57c-58c, with a CPU temp of 46c.

 
Those temps seem pretty high for 3.0 GHz. I was able to run at 3.0 GHz with stock voltage. For 3.3 GHz, I needed 1.30 volts. Under Prime95, CoreTemp reported 2 cores (0 and 3) at 50 C. The other two were were at 55 C. To run at 3.6 GHz, I needed to raise Vcore to 1.45 volts (droops to 1.40 volts). Temps are 60 - 65 C.

MB is a GA-EP35-DS3P. HSF is a TRUE/ S-Flex fan combo.

Sailor is right. By today's standards, Zalman's just aren't that good. But they sure are pretty. I have a 9500 (eBay) on my knick knack shelf. Attracts a fair amount of attention when I have company over.
 
Yep, the TRUE's heatpipes kept on touching the northbridge heatsink cooling blades and wouldn't allow the TRUE to seat properly. I was bummed.
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom