OC'ing Q6600 some questions

xringx

Distinguished
May 2, 2007
475
0
18,790
Ok, just got my q6600 (GO) and working my OC on it. I have some questions...

Specs 1st.

Q6600 Not lapped
GA-965P-DS3 rev 1.0 mobo
4GB Crucial Ballastix
Vista Ultimate 64bit
TRUE lapped for heatsink

1. Does speed fan work correctly? Its reporting my temps full load at 3.2 (356x9) at 30c,30c,24c,24c @ 1.32 vcore. Seems really low on the temp side

2. I am having trouble getting anything higher. I have tried 3.3 at 8multi and 9multi upto a 1.40 vcore and nothing is stable. Any recommendations?

I dont feel comfortable pushing higher vcore. Also, I am not all the experience with vcore drop, but when running OCCT I notice that the vcore per speed fan drops alot (down to like 1.28) Is this a potential problem? Should I raise the vcore in bios to compensate?

Thanks for your advice and help
 

xringx

Distinguished
May 2, 2007
475
0
18,790
Ok, I dl"ed coretemp, and it reports different temps. Full load on 4 cores reports 48, 48, 45,45 which seems better....

VID per core temp is 1.30. I would swear that my vcore is at least 1 tick higher in bios.. Need to check it out
 

Lupiron

Distinguished
Feb 9, 2008
1,711
0
19,780
I think speed fan was one of the progies that needed to be calibrated on some computers, by 15 c or whatever? Though the TRUE will keep things nice and cool!

Yeah, thats what is prolly going on, if you have a large droop, you need more juice, just for when it is fully loaded. I have a p5k-e that has a vdroop of .02 normally, and the 780i board I have starts out at .06, and gets to .08 in normal 1.4500 type ranges. You need more core voltage, just because it drops so damn low, and that does suck!

As far as VCore goes, its rated at 1.5 volts max, and I have my g0 6600 at 1.50 volts, at 3.8. Its still cool with a TRUE though, low 70's when testing, but daily, I dont see a core at 60c.

I think the rumor is that as long as you stay within spec, and near 65c on the max core temps, it'll last a decent amount of time.

got me.

--Lupi
 

billiardicus

Distinguished
Mar 2, 2008
186
0
18,680
Assuming your memory is not causing the instability at higher cpu speeds, and increasing vcore doesn't improve stability, consider increasing the FSB and or northbridge voltage a little. My Q6600 (go) wouldn't OC past 3.3ghz until I bumped the FSB and NB voltages
 

xringx

Distinguished
May 2, 2007
475
0
18,790
Well, I tried that. Here is the thing. I upped both.. and i couldnt even boot into 3.2, no mater the voltage (1.50) even...

need to play with it more
 

xringx

Distinguished
May 2, 2007
475
0
18,790
Ok, things I have learned.

1. with the 8multi, i dont even get into the windows boot up screen before a restart.

2. With MCH or FSB voltages, even with 1 little tick up, wont boot into what was a previously stable boot.

3. At 3.4 (9x378) no matter the voltage, I get a restart on windows boot.

4. Highest boot is 3.348 (9x372) at a bios vcore of 1.3375.

I am beginning to think that my MB might be limiting me... It is abit of an older MB....
 

Lupiron

Distinguished
Feb 9, 2008
1,711
0
19,780
I scoped it out and it does look like an older board. With the 8 x multi and a higher FSB it doesnt boot? Or are you still trying a low FSB? Maybe the board simply cant handle a fast FSB.

It appears that all the newer boards are near 500 or better. Maybe that board was produced before everyone decided a high FSB is overall better for speed? Who knows!

You have a TRUE with a g0 chip, so you should easily be able to OC higher than that, and your ram is decent.

Have you tried playing with the multi to keep the speed as near to 2.4 for ya can, and leaving the memory at its SPD speed, prolly 5-5-5-15 @ 1.8, but that way you know its not the RAM! Mean while just keep upping the FSB until it fails to boot. When you do that, you can then use a higher multiplier and vcore and see if it again fails at the same FSB or within a few points. Do that a couple times with different settings and it becomes pretty clear that its the mobo's FSB limiting you. Just try to keep everything as near to stock as ya can, while adjusting what you need to for the higher FSB.

Then again... who knows?

--Lupi
 

xringx

Distinguished
May 2, 2007
475
0
18,790
As for the FSB, this MB can more than handle it, my last CPU (see spec) I can get 7x485.. So its not he FSB.....

I havent tried memory voltage of 1.8. Had it at 2.2 which is spec for timings of 4-4-4-12. But I cant see that being the problem, but will try it.

I may exchange it cpu...
 

Lupiron

Distinguished
Feb 9, 2008
1,711
0
19,780
Oh, the quad worked, but the e6300 isnt reaching faster speeds? Then you know its not the board, and its always best to test it OC'd with the rest on stock until you get the processor stable, as the rest is easier, but may for whatever reason interfere with the OC in general, so best to remove the doubt!

I'm not familiar with the voltages on the E series, so I cant help ya there! Maybe you can probe Evil Oni's brain for information! He's like a walking encyclopedia!

--Lupi
 

xringx

Distinguished
May 2, 2007
475
0
18,790
you mis understand me. I was able to reach a much higher FSB on the e6300 no problem.. but cant do it with this cpu. With the Q6600, I tried 7x458 and it wont even boot.

So i must assume its a combo of the cpu in conjuntion with the MB. I probably would have better results with a new generation of MB...
 

Lupiron

Distinguished
Feb 9, 2008
1,711
0
19,780
Oh! Yeah, that is odd, even for a high VID like 1.3000, seen a bunch of em around lately and they usually get to 3.6 without to much fuss. So yeah, I dont suppose you have a way of testing it in a friends board or some such? Or anything outside yours?

Not much else to do if you've tried disabling everything you dont use that comes with the board, like lan ports, serial ports, all CPU options, all spread spectrum options.. and if you've already tried a full range of voltages all the way up high... time to at least test it in another board.

For instance, Frys has a wonderful 30 day no questions return policy on their p5k-e wifi! (Or any other... sad to say it, but thats basically free testing, as long as you pay for gas, and let them hold the 150$ in collateral! LOL!)
 

Perp

Distinguished
Feb 27, 2008
144
0
18,680



You realize quads will produce more problems for higher FSB speeds right?

Hence why low multiplier 45nm quads aren't good overclockers. Fortunately for you the q6600 has a 9x multi which should allow you a fairly easy 3.6ghz. Any speeds you can get above that are much more of a challenge.

Also check out the extreme series overclocks on LN. Most of those guys are still only using 450ish fsb with high multipliers.
 

xringx

Distinguished
May 2, 2007
475
0
18,790
I am catching on to that. But what is wierd, I cant get past 3.2 stable. No matter the steps I take.

I am assuming that its the older MB holding it back some and not the CPU...

An x38 mobo with crossfire is in the works, but some many months away......
 

Perp

Distinguished
Feb 27, 2008
144
0
18,680
No supprise to me.

There are frequency/voltage issues between the FSB and the CPU that go way over my head tbh.

For example my q6600. It was stable at 3.0ghz then again at 3.6ghz, all with 9x multiplier. Funny thing is it won't post with 400fsb and 8x multiplier, yet it will with 9x.