Q6600 OC Questions

pieplough

pieplough

pieplough

Soldato
Joined
15 Jul 2008
Posts
3,636
Location
Glasgow
Evening, having recently procured a Q6600 SLACR G0 stepping chip for 75 notes im looking forward to seeing how high it will clock.

So anyone have experience of clocking it on a Abit ip35p + 4 gig of kingston HyperX ram? Cooling is a ageing Scythe NINJA.

Thanks :)
 
I'm on an Abit IP35 Pro XE with a Q6600, so should be broadly similar to your board.

I was stable at 3.2GHz (356x9, 1.3625Vcore) for a couple of years, but since I swapped out my old Scythe Zipang for a Megahalems the other week, I've been trying to see how far I can push on.

Not very so far! Currently at 3.32GHz, and seem to be 9 hours Prime blend stable but get odd false "success" messages in IBT before it should really have finished.

My CPU VID is 1.3175, so I'm not expecting great things myself, but with a Megahalems, a Spot Cool on my MOSFETs, and a HAF932 keeping my temps well under control, I'm gunning for 3.4GHz.

What do you want to know, exactly? I've not finished playing with my settings yet, and as I've hit a wall a 3.32 as far as general aimless tinkering will go, I need to get systematic, as follows:

1. Lower CPU multi to lowest value.
2. Keep increasing FSB until I don't pass 8 runs of IBT. Knock up MCH 1.25 (Northbridge) volt one step at a time and repeat until I seem to hit the FSB wall for the mobo. Also try single-step ICH 1.05 voltage increases if adding NB volts doesn't stabilise. I now know that the mobo FSB setting won't be holding my clock back for any FSB under these values with these volts.
3. Put CPU multi back to 9, keep NB voltage at settings for step 2. Start from a known stable CPU clock and Vcore (e.g. 9x356 for me), and increase FSB in increments of 5 or so.
4. If there are any stability issues, knock up Vcore a notch. If a couple of notches don't work for any point below mobo FSB wall, try increasing CPU VTT as well, one step at a time.

Oh, yeah, I would keep my RAM locked down to a low speed on the divider as well throughout this, and then try to increase that once I had everything else stable. Keep Vcore below 1.5v, don't go crazy with NB volts, and watch your temps!

This method is a pain, but once you get nearer the limits of your CPU/mobo, you need to figure out which factor is holding you back. From stock, a lot of people have some success with whacking FSB up to 333 (for 3.0GHz) and just bumping Vcore a little bit until they're stable. I got to 3.2GHz originally with that method. Depends how far you want to go. Me, I'm on a mission for 3.4GHz...
 
You should be able to get 3.6 so long as your cooling is good and the chip is capable of it, which to be fair, a big percentage are.

What you need to do with these boards is keep an eye on PWM temps and if possible do the heatpipe mod which involves removing the heatpipe and replacing the thermal rubbish with decent paste and bolting it back down with nuts and bolts rather than the push pins. This makes a BIG difference to PWM temps.

You can also make a big difference by getting some active cooling over the PWMs. Pop a 60mm fan or something over there. This alone can knock up to 20c or more off the temps when loaded. Search for my Fortress log to see mine.

The other thing to watch is that these boards are notorious for Vdroop. What you set vcore to in the bios can be actually anything up to 0.1 volt or more off when loaded.

These are my voltage settings for 3.6....

Vcore of 1.52 bios (1.42 - 1.43 under load according to uGuru)
VTT at 1.23
MCH at 1.29
the others are default
 
Oooh. Just looked up your Fortress build myself. Very inventive solution to PWM cooling! Looks like your results with that are about 10C better than my Spot Cool. Might consider something similar...

And, yeah, the vdroop on the IP35s is a PITA as well.
 
Thanks strumpusplunket and Diggsy for your detailed answers. I was hoping my CPU was going to arrive today but looks like it will be tomor, so all things going well ill get it installed and start clocking on saturday.

Ill see how far this board and cooler will let me go. I may end up buying a better Mobo if it is a limiting factor. Im sure there are good overclocking s775 boards to be had fairly cheap now.

I will probably be a bit more systematic and follow your steps strumpusplunket, my current oc was just following a guide then cranked it as high as i was comfy. Im hoping for at least 3.4 GHz, glad im not being to optimistic.
 
The Abit IP35 Pro is a great clocker, you just have to keep it cool, I run mine at 438 FSB with 8x multi for 3.5ghz without trouble, but to get it over 3.6ghz is going to take better cooling of the PWM areas as it quickly hits 102c doing IBT at 3.6 or above, I have actually booted into windows @3.7 but the system rebooted durning the IBT test as the PWM got over 102c...
 
Thats good to hear Tweak, im sure ive read somewhere that 9x can be a bit flakey? Im looking forward to seeing how high mine will go, got some MX-3 coming tomor from OcUK also.

Will be in a CM 690 with 3x 120 intake and 3x 120 exhaust so hope it will be cool enough in there.
 
Im not sure, but I think my Ram is causing my problems as i have two types of OCZ, 2x 2gb 5,6,6,18 @2.1 and 2x1gb 5,5,5,15 @2.3v, both PC2 8500.

I have manually set the timings in bios to 5,6,6,18 @2.3v to cover both and this works for 438x8 and ram @1051mhz, but if i go with 389x9 and ram at 1:1 = 778mhz, the system is unstable, even though I know the chip can easily do 3.5ghz @438fsb, and the ram works fine at 1051mhz???

I'll be looking to add a few small server fans to the pwm tonight and if i can get the temps down, I will remove the 2x1gb and start OC all over again to get as high as i can stable, then will put the 2x1gb back in to see if it causes instability.... hope not as i like having 6gb, doubt i need it though LOL.
 
Well added a 40/20mm fan to the northbridge, but dont have room to get one on the VRM cooler due to teh size of my CPU cooler :(
so PWM temps still getting far too high, looks like im stuck @3.5ghz :(
 
pieplough said:
Thats good to hear Tweak, im sure ive read somewhere that 9x can be a bit flakey? Im looking forward to seeing how high mine will go, got some MX-3 coming tomor from OcUK also.

Will be in a CM 690 with 3x 120 intake and 3x 120 exhaust so hope it will be cool enough in there.
Click to expand...

9x is fine. I run at 9x 400, but it is equally as happy to run 8x 450.

I plan on trying some new cooling at some point to see how high it will go.
 
Started clocking my chip today. Here is a pic of it during prime95

Untitled-9.jpg


9x333 nothing else changed. Tried to boot at 9x400 but wouldnt post and had to reset cmos.

Firstly, are these temps fine and secondly what should i set my ram at? 1:1 00 etc?
 
Ok, following some of diggsys advice in another thread ive got it running at 3.6

Settings:

fsb - 400
multi - 9
dram - 1:1:25
cpu - 1.52
cpu VTT - 1.23
mch - 1.29
ich - 1.5
cpugtl - 67%

Just cant seem to get it prime stable. It boots into windows fine and my idle temps are 33-35 load hasnt gone above 60 until it resets. Any ideas?
 
What are your actual volts when at full load? This board is notorious for a lot of vdroop. You may just need a tiny bit more vcore to get stability.

EDIT - In fact I see your CPU has a VID of 1.325, so it is quite likely you will need just a little more than mine (which is 1.3v VID) to get Prime stable.

EDIT2 - are you sure load hasn't gone above 60c at 3.6? Your screenshot at 3.0 above shows max temps at 58. Make sure you check this.

Also keep an eye on PWM temps, what are they at load?
 
Last edited:
Hey man, thanks for the reply.

Volts in cpu-z whilst running prime are 1.52 and in the bios they were set at 1.58. I tried increasing another notch but it was to high and as soon as prime started it reset.

The highest i seen the temps before reset was 62.

I tried to check my NB temps with everest but couldnt see it? Im not sure if the abit ip35p (not pro :() has NB sensors?

Im sure this chip will do 3.6 on air, it always boots in fine at that speed only prime kills it. I tried upping the MCH slightly and tried running the ram at 800 instead of 1000 but still no joy :(

Appreciate your time and help.
 
I'd suggest downloading CPUID Hardware Monitor and checking your PWM temperature as well as your core temp.

The PWMs on this board get very hot, as a few of us have said, but you don't mention if you've monitored them. It's quite possible that your PC shut down because they went over the shut-down threshold. With CPUID HW you can keep an eye on them. I'd imagine that with 1.52vcore you're heating them up very quickly.
 
As above really, use HW Monitor to keep an eye on the PWM temp and also CPU voltage during load.

You're right too, there is no sensor that I am aware of that measures the NB temp on these boards, but you can certainly help by adding some active cooling on there. I run 2x 40mm fans on mine.

I wouldn't advise running at 1.52v real for any length of time, unless you are under water or better cooling. To be honest your chip shouldn't really need that much for 3.6

I would suggest you reduce the vcore to 1.52v bios and try bumping the MCH to 1.35 or even 1.4 and see if that helps.
 
Looks like it's labelled AUXTIN on yours for some reason. I reckon that's PWM, though. Check those temps under heavy load and you should see it shoot up.

EDIT: yes, mine too has a lot more info. I don't have an AUXTIN temp, but I do that SYS, CPU and PWM together in that same place, which is why I thought that was your PWM. Diggsy has that temp and PWM, though, so I guess maybe not...
 
You must log in or register to reply here.
Back
Top Bottom