Exploring auto voltage implications with an old Sandy

[BonzaiDuck]

Tell the truth -- I haven't kept up (firsthand) with anything beyond my 2600K. All my overclocking with the CPU was finished before November, '11. I OC'd the RAM last year, but it created an infrequent instability. Haven't really spent as much time with it as when it was built.

On the P8Z68 board, there are several levels of "auto." Changing to "offset" mode removes one level; changing offset to a fixed value removes another; changing "Extra voltage in Turbo" from "auto" to a fixed value removes a third. But there's one "level" remaining.

I wanted to see how my voltages looked just by dropping the multiplier from 46 to 44. The EIST idle voltage is still the same. The loaded Prime95 voltage is now 0.02V lower than it was, and the transition load-to-idle peak voltage -- 1.36+ @ 46 -- is now 1.34+ @ 44.

Looking at my 2011 notes, I see that a provisionally successful IBT and Prime95 test at 45 was 0.02V lower across the same range, but those were preliminary, and I'd adusted my "4.6 preliminary" upward for consistent maximum GFLOPs.

Used to have this idea that various intermediate clock multiplier settings should be fed by the voltages of the proven tests. Now -- I don't think so.

It would make just as much sense to change only the multiplier and leave the voltage settings alone.

I know . . . . "What's the big deal?" I just think it's a good thing to know. . . .

 

[FalseChristian]

I use Offset Voltage on my i5 2500K at 4.765GHz. The voltage drops to 1.45v when under load and remains at 1.50v idle.

 

[Deders]

I know with the latest generation of Asus boards, the engineers have dona a lot to fine tune the automatic settings, an experienced friend of mine managed to get better voltages with the auto tune than with his own settings.

 

[BonzaiDuck]

I use Offset Voltage on my i5 2500K at 4.765GHz. The voltage drops to 1.45v when under load and remains at 1.50v idle.

Is it still tip-top and alive? IDontCare -- one of our beatified gurus (he earns it) -- is optimistic about the safe or operable voltage range and processor longevity/integrity on the SB K chips. But I think I could get to 4.70Ghz with all measurable voltages under 1.40V -- maybe even 1.39 or 1.38. Also -- sounds to me from your decimal places that you tweaked the bCLK somewhere between 103 and 105 Mhz.

Truth is -- I did that, too. In fact -- I used the voltage settings for a 4.6+Ghz OC to begin tuning the 4.60Ghz speed. I just decided the voltage headroom was good to finish the final adjustments, and the bCLK wasn't really worth it to me.

I know with the latest generation of Asus boards, the engineers have dona a lot to fine tune the automatic settings, an experienced friend of mine managed to get better voltages with the auto tune than with his own settings.

Well, I think the feature in the "AI Tweaker" BIOS menu is called "OC Tuner." The first step of "rave success" for a 2600K chip is the 4.4 or 4.44 Ghz result. I seem to remember that the final voltage seemed too high. I think it was over 1.40V, but that's a recollection more than 2 years old.