ASUS AMD Beta BIOS Releases

[chew*]

Maybe Hexacore CPUs will have a more heavy overvolting.
My 955 set 1.4625v in BIOS, CPU-Z shows 1.476v on idle and 1.488v on load, same as previous BIOS.

load as in prime 95 or just a game?

I'm set at about 1.4375 for example right now to compensate doing some testing in prime 95, Load = 1.488-1.50. I use LLC only to keep voltage more consistent, it bounces around quite a bit more off.

Target is 1.475 which I will hit with these settings IDLE.

It may have more to do with the way turbo is implemented on chip. I like control and don't want my chip doing stuff I don't tell it to do especially when doing reviews, that said turbo is off. What could be happening is turbo is told to be off multi wise but the cpu and or microcode isn't smart enough to say hey we don't need to ramp voltage now becasue turbo is off. Thats a very logical explanation why instead of having some droop voltage ramps up even more when loaded as that is the way turbo was designed to work.

It's not a big deal LLC percentages will allow closer to accurate set voltages for all parties.

[Royalk]

load as in prime 95 or just a game?

I'm set at about 1.4375 for example right now to compensate doing some testing in prime 95, Load = 1.488-1.50. I use LLC only to keep voltage more consistent, it bounces around quite a bit more off.

Target is 1.475 which I will hit with these settings IDLE.

It may have more to do with the way turbo is implemented on chip. I like control and don't want my chip doing stuff I don't tell it to do especially when doing reviews, that said turbo is off. What could be happening is turbo is told to be off multi wise but the cpu and or microcode isn't smart enough to say hey we don't need to ramp voltage now becasue turbo is off. Thats a very logical explanation why instead of having some droop voltage ramps up even more when loaded as that is the way turbo was designed to work.

It's not a big deal LLC percentages will allow closer to accurate set voltages for all parties.

Load on Prime 95. Disabling LLC may cause a ~0.06v vdroop for my x4 955.

Changing to 1055T and the same board, LLC overvolt goes higher. Set 1.4125v for 4GHz and shows 1.452-1.464v on CPU-Z under Prime 95 load. Similarly as yours.

I'm struggling if CPU-Z voltage value is a bit higher than actual on M4A89TD? I have measured vcore use a voltmeter on C4F, and it is always ~0.02v lower than what CPU-Z showed (1.4125v 4GHz setted in BIOS as M4A89TD, 1.438-1.440v measured, 1.452-1.464v on CPU-Z under Prime 95 load), I doubt this because I'm not sure if C4F's LLC level is the same as M4A89TD.

[chew*]

Load on Prime 95. Disabling LLC may cause a ~0.06v vdroop for my x4 955.

Changing to 1055T and the same board, LLC overvolt goes higher. Set 1.4125v for 4GHz and shows 1.452-1.464v on CPU-Z under Prime 95 load. Similarly as yours.

I'm struggling if CPU-Z voltage value is a bit higher than actual on M4A89TD? I have measured vcore use a voltmeter on C4F, and it is always ~0.02v lower than what CPU-Z showed (1.4125v 4GHz setted in BIOS as M4A89TD, 1.438-1.440v measured, 1.452-1.464v on CPU-Z under Prime 95 load), I doubt this because I'm not sure if C4F's LLC level is the same as M4A89TD.

I've measured, it's preety accurate.