Is this correct or not - Please can someone respond - Thanks

[nicksort]

Can someone tell me if I am doing this correctly or not?

I have another thread about what I have done so far but I did something different today.
I have tried to read as many threads about overclocking but I am still not grasping it all the way so i have my doubts at what I am doing.

E6400 P5B m/b 667 DDR ram

so I have been moving things around yesterday and managed to get 3.0 with changes to the FSB only and it was around 66-68C.
Today I changed and started at 2.92 and dropped the cpu V to 1.28 and it ran real good with 44-46C
Next the tried 3.0 and kept the cpu V at 1.28 and a NO windows boot up.
I then changed the cpu V to 1.30 and she booted right up and running smoothly and the temps are at 50-52C. I think this is a big improvment from yesterday.
I still need to buy a different than stock fan cooler so I am worried a bit.

My question is, AM I going in the right or wrong direction?

Thanks.....

 

[happy medium]

Sounds like you going in the right direction to me.

 

[Dopekitten]

Yes you are, but be sure to use Prime95 to test for stability, i don't know if you are or not already.

 

[Denithor]

Run the lowest voltage you can at the speed you want to hit to keep temps low. See what your temps are like with orthos (or another stability test) running with the stock cooler.

If you're only shooting for 3GHz you should be fine on the stock cooler (I ran my e6400 @ 3GHz for over a year with stock cooling with no problems at all, have since upgraded to e8400--and don't really see any difference). Keep your temps to 60C or lower and you should be gtg.

 

[soccerballtux]

You can hit 73.2C and be fine.

 

[Martimus]

Originally posted by: soccerballtux
You can hit 73.2C and be fine.

I kind of doubt that the thermocouples on the CPU are that accurate.

 

[Idontcare]

Originally posted by: Martimus

Originally posted by: soccerballtux
You can hit 73.2C and be fine.

I kind of doubt that the thermocouples on the CPU are that accurate.

heh heh, well he didn't exactly say you would know when you are hitting that temp...soccerball just said you can (you just might not realize you are) and you'll be fine

 

[soccerballtux]

Originally posted by: Idontcare

Originally posted by: Martimus

Originally posted by: soccerballtux
You can hit 73.2C and be fine.

I kind of doubt that the thermocouples on the CPU are that accurate.

heh heh, well he didn't exactly say you would know when you are hitting that temp...soccerball just said you can (you just might not realize you are) and you'll be fine

Intel guarantees accuracy for thermal cutoff temps (IE, when you're getting from Tjunction 73.2C, to 90C). For lower temps (idle) they may be a couple degrees off, but when you get up there they're spot on. My guess is this is why my 2nd core is always cooler than my 1st core...until I hit about 70C where they even out. I lapped them pretty well I thought, and I don't think it's too much thermal compound because I've reseated multiple times. Besides, it would be weird (if indeed I lapped/applied thermal paste poorly) that at 70C the CPU die suddenly became a much better thermal dissipator.

My maximum temp is 74C in Orthos, it's usually much lower than that (gaming), and I've been fine (since January).

 

[nicksort]

Not to sound stupid but "What is Ortho"?????

 

[BonzaiDuck]

Originally posted by: nicksort

Not to sound stupid but "What is Ortho"?????

In Graysky's over-clocking sticky -- possibly also the sticky on stress-testing -- you'll find a couple of links to "Mersenne.org." The web-site was set up to enlist people for using their PCs to find new prime numbers with routines that do Fourier Transforms. We're not interested in the math here, but only in the fact that the programs tax the dickens out of a CPU.

So appendages and revisions arose in the PRIME95 program devoted exclusively to stress-testing.

With the advent of dual-core processors, it became necessary to have two PRIME95 instances running to properly load the CPU for testing. So somebody revised the program and created ORTHOS.

When the quad-cores became available, there was then a need for a four-instance PRIME95 program, so someone simply invented one which detects the number of cores and runs that many instances of PRIME95.