Stupid question - why do intel voltages vary?

[davidm71]

I have an E6600 and some reviews have it at 1.25volts as the default voltage. On my motherboard the minimal voltage setting is 1.325 volts. Is that because some e6600s didn't require a lot of volts and later some of the rejects needed more volts to sell as an e6600?

Thanks.

 

[BinarySynapse]

The silicon wafer that's used to make the processor isn't perfectly uniform. So each chip is going to behave differently and have different requirements to get them to work as they're supposed to.

 

[davidm71]

Then does that mean I got a poor overclocker if one takes 1.25 and mine takes 1.325? Also on the box it says 1.35v Max, and intels website has, according to the processor serial id, spec'd at 0.8-1.5v. Why would there be a disparity? Guess that means I can run it at 1.42 volts?


Thanks.

 

[BinarySynapse]

Then does that mean I got a poor overclocker if one takes 1.25 and mine takes 1.325? Also on the box it says 1.35v Max, and intels website has, according to the processor serial id, spec'd at 0.8-1.5v. Why would there be a disparity? Guess that means I can run it at 1.42 volts?


Thanks.

No it wouldn't necessarily be a bad overclocker. Lower voltage chips do tend to over clock better, but it's certainly not guaranteed. There's an nearly endless number of variables that can affect overclocking, most of which we can't even see or know about. A low voltage chip can hit a brick wall while a higher voltage one reaches the sky.

As far as the voltage disparity is concerned, I have no idea.

 

[Deleted member 184142]

No it wouldn't necessarily be a bad overclocker. Lower voltage chips do tend to over clock better, but it's certainly not guaranteed. There's an nearly endless number of variables that can affect overclocking, most of which we can't even see or know about. A low voltage chip can hit a brick wall while a higher voltage one reaches the sky.

+1

Also, 65nm chips are "safe" to 1.5v. Also, have you tried testing to see if your chip is stable at a lower voltage? My Q9550 on auto was set at 1.187v but I could undervolt it all the way to 3GHz (2.83GHz stock). The auto voltage is not always the lowest it will run at stock speeds.

 

[davidm71]

+1

Also, 65nm chips are "safe" to 1.5v. Also, have you tried testing to see if your chip is stable at a lower voltage? My Q9550 on auto was set at 1.187v but I could undervolt it all the way to 3GHz (2.83GHz stock). The auto voltage is not always the lowest it will run at stock speeds.

For the most part I ran my 2.4ghz E6600 at 2.4 to 2.7ghz for the last couple years with my 8500 ram at 800mhz tight timings. Now I'm willing to take more chances as I am almost ready for a new board by 2010 or late 2009. Stock voltage was only good for up to 2.7. So I tried raising the CPU freq all the way to 3.3ghz (370x9) and it BSOD upon boot even with 1.355v. So lowered it down to 3.15ghz and it loaded windows but orthos blue screened in five minutes. Dropped the multiplier to 8 and raised fsb to 355 at 2.85ghz and it ran orthos for hours. Then I upped the ram to 1067 at its rated 2.25v and up'd the nb to 1.4. Still stable.

Anyhow, If I lowered the multiplier so that my fsb was at 360 and the cpu was stock it was stable. Wanted to rule out my motherboard as a limiting factor. So then at 1.42 vcore and nb,sb, and hypertransport up one notch I was able to boot, and run orthos stable for over an hour at 3.195 ghz.

Heres my data:

300x9 : 1.335v
333x9: 1.365v
355x9 : 1.42v nb, sb, hypertransport +1 notch.

My temps are up: each core is around 37-40 deg C under load and the PWMs are 10 degrees warmer around 50-54 degrees C. But it seems initially stable. Lets see how it does running Farcry2 in SLI!

Thanks.