[Spraynwipe]

Hey guys, my setup is q6600 G0, Asus blitz formula, 2x http://www.ocztechnology.com/product...c_dual_channel as my RAM (DDR2-800).

I am running at 3.25GHz with the following settings
Multi: 9x
FSB strap to north bridge: 400MHz
FSB: 360
DRAM Frequency: 959MHz
DRAM voltage: 2.10V

RAM Timings: 5-6-6-18
FSB:RAM ratio is 16:12, or 4:3

CPU voltage: 1.475 perhaps a bit high, but can't go any lower for some reason and I'm worried it's because of the RAM settings.

Is the 4:3 FSB:RAM ratio too high, and what about my other settings? I'm still relatively new to OCing, and recently had some BSOD crashes related to overclocking. However, in the past, I never had an issue running 3.25GHz with lower voltage (around 1.4V) but now I can't run 3.25GHz below 1.475V and I'm starting to think it's something to do with the RAM settings, hence this post. I don't think it is hardware failure (even though the BSOD I was getting indicated so) because it runs fine at stock.

Really would appreciate any help or insight into my problem, thanks!

 

[Ovlovian]

With temps under control your CPU Voltage is fine that high.

Personally I would set my ram timings at 2:1 considering it's 800 and it would effectively be 720 at your FSB

Shouldn't you be able to run at 5-5-5-18?

Your Ram should be run at 1.8v

 

[Spraynwipe]

Okay thanks for the info, I'll be changing to 8x400 (3.2GHz is more than enough), ram to 1.8V, timings at 5-5-5-18. In this case, what should my RAM frequency be set to?

Also going to drop the cpu voltage down to 1.45, does that seem a little excessive for 3.2GHz?

 

[Spraynwipe]

I've done those settings, set the RAM frequency to 800 and going to try some stability tests now. Any further help or advice is welcome

 

[Ovlovian]

Understanding why I'm telling you certain settings is quite a bit more important than the settings themselves. Voltages are something you raise to enable you to run higher clocks. Until you verify that it is what's holding you back you should adjust it at all. Voltage is what kills components typically. For example, you were running ram at relaxed timings and a lowered speed but at much higher voltage than stock. Ram timings (those 4 numbers with dashes) translate to quicker ram the lower they are set.
In so far as your chip is concerned id shoot for closer to 1.38v if possible. A key point with overclocking I've learned through car work. Change one thing at a time and make small changes.

 

[neoresin]

excuse me if i might jump in here guys and say something real quick. as i understand it, intel's 'safe' VID (which as i understand is equivalent to vcore) for the q6600 g0 is 0.85-1.5v:

http://processorfinder.intel.com/det...px?sSpec=SLACR

and reading intel's documentation on the subject tells us that as long as we keep the VID/Vcore between these levels, we can expect long term reliability and operation (within temperature limits as well, of course - mid-60s C on the cores).

i must say my disclaimer right now though: understand i'm just getting into this, doing my first oc just this week. i've done a fair bit of reading though (hopefully enough to be my fair share), so don't think this is coming from nowhere. i've got my q6600 g0 set to a VID in the bios of 1.5375 V right now, which translates (after load-line calibration and vdroop, as i understand it) to 1.504 V in CPU-Z, which i expect to be within limits set by intel and operational for a while (but what do i know).

anyway, i just wanted to jump in and say my two cents. by all means, if someone has something to say in reply, don't hesitate to either strike me down (lol) or perhaps give some new information. i, for one, would appreciate it.

 

[Ovlovian]

Some chips can run higher clocks than others at a given Vcore. The lower the Vcore, the less heat is generated due to resistance (Because of the decreased consumption of power) and the better it is for the chip. All chips are not created equally. Exceeding the VID recommended is at your own risk and not something I'd do on a system that I'd like to last for a long time/24-7 folding.

 

[neoresin]

wow, ok sorry. didn't know this thread was dead.

but i guess that's what i'm asking: would the VID be considered equal to the Vcore? or would they be two separate entities?

also: wouldn't all chips of the same model (hopefully) be considered created equally? i mean, i know there's inherent risks posed when a chip is overclocked past its manufacturer's limits. but its pretty safe to assume that they've done extensive testing on these things, and as i've read, the core2quads can take a beating and still keep on ticking.

 

[Ovlovian]

If they were created equal binning wouldn't be necessary. Take the Xeon vs i7 chips. W3530 is just a 930 i7 that was tested and proved it was a great example of that chip. Maybe it just did stock clocks at much lower voltage than expected. Maybe it OC'ed much higher than the others. Who knows. For whatever reason it was just a "Better" chip than others so they can sell it as a higher/different badging because they know it'll be more reliable. This is why people buy from the Xeon line.

 

[neoresin]

oh okay. this is a valuable piece of info. thanks.

so but, i guess that means you can never really compare your own experiences with a chip with others'. names mean nothing. but bins mean everything.

 

[Ovlovian]

Yes and no. You can use others as a frame of reference but you never really know till you try your own chip.