Intel Burn-in Test

So I was running my new Q9650 at 4.0 stable in Prime95... but when I tried the Intel Burn in Test, I got errors almost immediately. So, I backed it down some to 3.6 and still got errors... and then down some more to stock speeds and still errors.

So now I'm wondering if my board can even really get this chip stable... Intel Burn in test doesn't even pass at default settings.

The vdroop on my board is probably the culprit. When I was OCing, I set it at 1.39vcore, which reads at about 1.35 on cpu-z. The trouble is that it drops down to 1.27-28 when stress testing, making the system unstable.

Wondering if I should RMA the chip. Any ideas?

[Logical]

Try Linx burn in test and compare results. Maybe your system doesnt like intel burn in, also make sure your hard disk is inactive at the time of testing.

Originally Posted by TrixteH

Try Linx burn in test and compare results. Maybe your system doesnt like intel burn in, also make sure your hard disk is inactive at the time of testing.

Linx or linux? Link?

I prefer OCCT now.

Originally Posted by grey_1

I prefer OCCT now.

Thanks, I'll give it a shot.

[bob saget]

you are so capable methim

[Logical]

Originally Posted by methimpikehoses

Linx or linux? Link?

http://www.fileden.com/files/2008/12/29/2243633/LinX.7z

I believe its very similar to the intel burn in....linx makes use of all the core I7's cores including the logical cores.

[Feyy]

Tbh if you are 12 hours prime95 stable, (all cores 100%).

Then you are fine, no game will ever stress it harder.

Well, I can't get this chip stable at the clocks I want on my P35 board, which is too bad... I'm sure it would be fine on a P45 board. Oh well, I'm not buying a new mobo just to OC a little higher. If newegg charges me a restock fee, I'll put it up for sale here.

Looks like I'll go back to the old Q6600.

[bacon12]

It may just be a bad clocking chip as well Meth.

The vdrop and vdroops on this board is pretty bad... I set vcore at 1.43, it reads at 1.36, and then under load the whole thing drops to 1.29. Pretty much makes it impossible to maintain a good OC.

The Q6600 was better stability wise on this board...

[mailman2]

Originally Posted by methimpikehoses

So I was running my new Q9650 at 4.0 stable in Prime95... but when I tried the Intel Burn in Test, I got errors almost immediately. So, I backed it down some to 3.6 and still got errors... and then down some more to stock speeds and still errors.

So now I'm wondering if my board can even really get this chip stable... Intel Burn in test doesn't even pass at default settings.

The vdroop on my board is probably the culprit. When I was OCing, I set it at 1.39vcore, which reads at about 1.35 on cpu-z. The trouble is that it drops down to 1.27-28 when stress testing, making the system unstable.

Wondering if I should RMA the chip. Any ideas?

I've owned a bunch of 45nm chips, they are flakey at best. Last night I came home, ran a quick 20 pass LinX at my 24/7 OC of 4.0Ghz it failed 9 seconds. Immediately ran it again and passed twice more. Rebooted and passed, then failed twice. I have found similar results using Prime95. It will pass 12 hours, stop it, restart it and fails within a minute.

I've had 5 E8400s C1 and E0s and this Q9550. My 65nm chips, which I also owned a multitude of were either stable or they werent but there wasnt any random instability like the 45nm chips exhibit.

The 45nm are still faster and have more cache, get your 12 hours of prime95 then just use it. Trying to prove it's stable over and over is an adventure in insanity. I've had 6 different 775 motherboard, 4 kits of ram and about 10 CPUs in the last year. The amount of random instability the 45nm chip exhibit is frustrating. However, I've never had issues using it to encode or game, its just the stress tests that cause this to happen.