MIA "CPU Turbo" switch

[xXlAinXx]

Really really long story, by the way I was playing around with "tweaked stock" O.C gear setting and Id noticed that as u can see by the below image, CPU turbo, even it's properly working: there missing option for enable/disable ( at least here ) it after testing out G25\G28



I m unable to restore it's "default" position even after reprogramming with some old bioses, the only one thing which I do over is removing the original bios which was sleepy on ROM2 ( I must have one backup somewhere else but atm Im unable to find it ).
So dunno if is possible trigger back the condition for let's that happening, cause Im unable to double check. Anyway ROM1 G25, ROM2 G28 and/or viceversa (prob only G25 are needed on boths ROM).
About that, Im pretty sure the condition for it's functional operation like signal quality and thermal specification are satisfied, I said that cause the multi are locked to x21 over prime95 blend test.



I've already checked it's presence in the .bin file and it's where it must be



So, no real clue about that MIA "CPU Turbo" option. Except that I even notice that after one CMOS battery discarge, the "routine" set some "wrong" QPI (6.4) as default and current voltage begin set @ 1.31250, so not unusual instead pretty stedy but kinda wired even for one http://processorfinder.intel.com/det...px?sSpec=SLBCH.

[saaya]

sounds like a corrupted bios to me...
i heard of that before but never saw it myself...

reflash the bios and check if it pops up again?
or maybe you changed some cpu feature option and it somehow results in cpu turbo not showing up?
did you disable ppm or eist?

[xXlAinXx]

Aye aye captain! let's call it "fixed" I had to traffic a little more with battery discarge, and set force BIOS ROM 1 trought jumper.
Anyway the QPI is still begin set with (6.4) @ default

[drnip]

What's that second utility you have listed? Is it motherboard based software or other?

[saaya]

the first he showed a ss of was hwinfo32, the second is a bios editor, he was checking if the option was hidden or removed from bios somehow.

about qpi, check with cpuz or hwinfo32 in windows please, iirc it shows as 6.4 in bios but it does set it correctly according to what the cpu flags as default qpi link speed.
its something about the bios not recognizing it easily and possibly an award kernel bug or something, dont remember the details...
since it didnt actually affect the qpi speed when we checked, was quite a while ago we first noticed this, we decided to ignore this bug.

but even if it actually sets the qpi speed wrong at default, it doesnt really matter, you dont lose any performance whatsoever from what i saw when i tested.
even if you underclock bclock to 85mhz (yepp, br can do that ) and set the lowest qpi multiplier, there is still no drop in performance... qpi is quite a powerful bus, well it is essentially hyper transport, and we all know how ht made no diference when benching on a64.

[massman]

even if you underclock bclock to 85mhz (yepp, br can do that )

Can it do that stable? I tested a couple of X58's back in the day for underclocking and A LOT OF them had issues around 100MHz.

[xXlAinXx]

What's that second utility you have listed? Is it motherboard based software or other?

the first he showed a ss of was hwinfo32, the second is a bios editor, he was checking if the option was hidden or removed from bios somehow.

;D

about qpi, check with cpuz or hwinfo32 in windows please, iirc it shows as 6.4 in bios but it does set it correctly according to what the cpu flags as default qpi link speed.
its something about the bios not recognizing it easily and possibly an award kernel bug or something, dont remember the details...
since it didnt actually affect the qpi speed when we checked, was quite a while ago we first noticed this, we decided to ignore this bug.

but even if it actually sets the qpi speed wrong at default, it doesnt really matter, you dont lose any performance whatsoever from what i saw when i tested.
even if you underclock bclock to 85mhz (yepp, br can do that ) and set the lowest qpi multiplier, there is still no drop in performance... qpi is quite a powerful bus, well it is essentially hyper transport, and we all know how ht made no diference when benching on a64.

Eheh, with all the EEPROM scattered around in these time a mere CMOS reset will not enought, especially while passing by some real different firmware/bios revision. As we been with G2x series!
It's doesnt matter, as u said if u leave (6.4) while is in AUTO configuration as soon u'd get to the late POST some manager will let the bios tell to update its register at the chipset, by then u'll stick with (4.8) default :P
((So, it might be a hook rountines (?shadow memory?) that glitch by the first initialization.))

[saaya]

hehehe i didnt know your that skilled in bios coding lain
nice

[saaya]

Can it do that stable? I tested a couple of X58's back in the day for underclocking and A LOT OF them had issues around 100MHz.

yes, i think 86mhz was 3d stable
i asked the bios engineer to tweak low bclocks so we, and later you guys, the customers, could test high cpu multipliers on air

unfortunately max cpu multipliers never worked properly... its weird, the bios guys could get 40+ multis working but only when one core is enabled. with more cores they were stuck to 30 or 31 and they never figured out how to solve it... sad, cause asus and gigabyte figured that out even before the 1366 launch :/

[massman]

Kinda reminds me of the issues I had while testing the board under LN2: with 4 cores I could only run 29x or lower and for 31x I needed to disable two of the four cores