Can't get Q6600 over 3.25

[stevehookem]

I can't get my Q6600 stable over 3.25. If I go higher, Windows won't boot. I have raised the vcore to 1.4 already, I raised the FSB voltage to the highest setting that is still green (can't remember the setting now).

I have a eVGA 780i board and 4gb of Dominator 1066 RAM. Currently, the RAM is set at 5 5 5 15 @ 1066. I haven't messed with it.

Can I get to 3.4? What should I check? Real question: will I see any "real world" speed increase in FSX if I go from 3.25 to 3.4? It is stable now and I don't want to lose that.

[NickN]

stuck are you?

3.25?

what FSB are you running? 406?

8x406=3250Mhz or 3.25GHz

that would mean you are stuck on the 8 multiplier and that would suggest the CPU Properties in the BIOS are not set to allow the full 9 multiplier. Under the CPU menu, DISABLE everything like CE1, SpeedStep, Vanderpool, etc, etc, etc.

That will allow the CPU to run at 9x and at that point 400MHz will give you 3.6GHz

In order to run 3.6 you will most likely need to run Vcore at about 1.42-1.45v in which you better be sure your HSF is able to handle the heat! I dont know that for sure as all systems are different however that is typical for a Q66 @ 3.6GHz

If you dont have a cooler that will keep that CPU under 70c at that vcore under FULL OCCT load test then you best not run that


If you are already on the locked 9 multiplier and have confirmed that then the problem is stability of the chip in vcore or heat, motherboard in FSB or the memory settings with the STRAP/MEMSPED and timing for the FSB you are trying to run

[stevehookem]

Oops, way over my head now! haha

I am using a 9 multiplier x 361.

When running the Intel Burn Test at 100% load, max torture, my temp has never risen above 54

[NickN]

raise the vcore.. typically you need to be around 1.44v to hit 3.6Ghz

as long as the full load (4 core load) temp reading using a load test software like OCCT does not exceed 75c, your fine. Even if you hit 80c with the chip its not going to damage it however you dont want to run that high.

and you need to make sure you are reading the diode temp, not the processor case temp. Use CoreTemp to read Core0 through Core3 while stress testing

[stevehookem]

I watch core temp as the test runs. Core 0 and 1 have hit 54

[stevehookem]

I bumped the vcore and I am running the burn test at 3.33 now.

My temps look really good---not higher than 47

[NickN]

9x400 will be 3.6GHz and the faster the CPU, the better the FSX


47c max?
Hmmmmmm

[stevehookem]

Actually when I last ran the burn test 10 times on max, I hit 58

[NickN]

[quote]Actually when I last ran the burn test 10 times on max, I hit 58

[stevehookem]

I apologize in advance for being a noob! haha

I have looked all through the BIOS for something that says "strap" and don't see it.

I have an eVGA 780i with Phoenix BIOS. It has some "nForce" settings that I don't know how to set. Do you know this board and BIOS?

If possible, can you walk me through the different settings?

[NickN]

Sorry.. I dont do nvidia motherboards...

and unfortunately I dont have time to read your motherboard manual and try to work it out.. I just got finished going through clocks and setup with 2 others and I am way behind on my own GEX work

[fighter25]

Ok Steve, my 750i FTW board has an FSB hole from about 1400-1520Mhz
I finaly got my Q6600 to 3.45GHz after jumping that "hole"
I don't know if this would apply to your board, but i think it might.

[stevehookem]

I'm stable at 3.6 now! Max temp in CoreTemp is 62

[GabethePilot]

don't forget that the voltage you set in the BIOS is not the "actual" voltage going through the CPU.

I set my Q6600 to 1.475 in the BIOS but with droop it manages around 1.39 to 1.42 under stress....1.46 at idle. with temps at a max of 65 degrees under stress - it'll do for a couple of years I reckon !

My CPU is running at 8x425 with a 1:1 RAM ratio (850Mhz) and I also increased the FSB and Northbridge (MCH) voltage by an extra 0.1v each.

This can increase stability depending on your m/b and whether you increase the FSB or the multiplier to gain your target speed.

Try both, and start with higher voltages to get things stable, then reduce voltages gradually until you reach the minimum needed for your stable o/c.

All this start low and increase nonsense is a waste of time unless you are dealing with unknowns - with the core duos, you're not though ! You know that the Q6600 can cope with up to 1.5 volts in the BIOS....so start high and work down - it's a lot quicker.

[stevehookem]

Ha! Absolutely!

I spent so much time messing around with tiny jumps from the bottom. I finally went directly to 3.6, raised the voltage to 1.475 and it was stable the first try. Lowered the voltage one (maybe two, can't recall) steps down and there it stays. No crashes at all--knock on wood!