Overclocking Prime95 fail - what does it mean?

Ive ran a stress test using PRIME95 and I got this image...can someone tell me what it means, I've just installed a fan on my CPU cooler and I was running prime95 to test temps....




if it has failed because of the overclock - would more voltage fix it?

or

I'm messing about with 333 rated DDR2 (1066 on order)- is this the cause of the overclock fail?

 

What are the temps of the cpu? more voltage might fix it but more voltage will raise the temps a lot so if it is already overheating I would lower the FSB

Also, what hardware is it?
cpu, motherboard, etc.

 

edit

 

rig is

Corsiar 600t : Intel q9550@3.8 : Asus P5Q Deluxe : 4Gig DDR2 : Gigabyte 460GTX : 550w Corsair : Creative SB X-Fi : Win 7 64bit Ultimate

 

yeah could probably raise the voltage slighlty if ur max temp is only 65

 

prime now running at 1.3 on vcore

how long do I leave prime running to make sure it's stable?

 

24 hours seems to be accepted timespan.

 

I've used 12 hours as my benchmark for stability and not had any issues thus far. Do not be fooled by 1 or 2 hour stable runs, I have had fails past this period and bluescreens when gaming.

 

temps went to 74 on 2 of the cores so given that the thermal limit on these chips is 71 or so...I've had to knock the vcore down to 1.28....see how prime likes that

 

Prime stable for 24 hrs if used as a folding machine
6-12 hrs if gaming is more than enough

 

I've asked this before but... IntelBurn Test?

I've heard around on that place known as the internet that a 5 minute test on that is as good as 12 hours of prime...

Not sure this is true but 5 mins seems more reasonable then 12 hours.

 

Agreed.
I've moved from P95 to IBT. It seems to weed out stability issues much more quickly, pushing the temps higher as well.

 

Do I smell a bit-tech article coming on?

I know they have enough to do already but...

*Edit*

Apologise for hi-jack.

 

thanks for the tip, it used to really get me to have to do a 12hr test

 

also try realtemp to give you a better temp reading.
also i find that the linpac test in OCCT is quite good as well

 

I had prime95 fail in 18 hours. Go figure

 

probably means your one voltage step away from a stable system

 

I had a mate who's stress test failed at 23 hours 50 odd minutes. It's why a lot of people do 24 hour runs. 12 hours doesn't mean it's stable. But then how many can be bothered to leave their PC occupied for 24hours? It's SO inconvenient if it's your only computer.

 

And what you use it for...

Most people don't need their CPU at 100% for 23-24 hours (except folding or rendering or w/e).