APU Fuse Interpreter (AFI)

[Laatanza]

no go on my Trinity

[Wizerty]

Thanks a lot for your works, and for sharing.

 

What's the "best" chips ?

low leakage or high leakage (for LN2) ?

and scalar high or low ?

[The Stilt]

LSlowmotion:

 

I'll fix that issue with 5600K asap.

 

 

Thanks a lot for your works, and for sharing.

 

What's the "best" chips ?

low leakage or high leakage (for LN2) ?

and scalar high or low ?

 

High leakage parts are a must for high frequencies.

Otherwise you will run into a voltage wall.

 

The best parts for high frequency have ultra high to extreme leakage and ultra low (<1980) leakage scalar.

Lower the leakage scalar value, higher the silicon quality (despite the leakage).

 

In my personal experience:

 

1.3375V = 7.6GHz

1.3125V = 7.8GHz

1.2875V = 8.0GHz

1.2750V = 8.2GHz(+)

1.2500V = ?.?GHz

Edited October 14, 2013 by The Stilt

[ObscureParadox]

Seriously tempted to have a go at this when I set up my FM2.

 

@The Stilt you sir are truly a legend!

[NoMS]

A10-5800K:

 

Same issue of LSlowmotion 5600K.

[flanker]

Seems not much bad CPU...lol. Not always I have lucky hand for chip (or its bug? All Trinitys here seems very bad)

 

[mekduk]

Hi all,

I am new here and really admire what Stilt has done for the APU community.

 

I have similar undeclared variables issue running on Windows 8, ASRock A85X Extreme 6, A6-6400K. May I know if this processor going to be supported in the future?

 

The program has been running smoothly on my A10-6800K though with the result as attached. Anyway, the leakage scalar is pretty low on this CPU but the rest of the value seems to be the middle of the road. May I know if this is going to be a good candidate for overclocking on air cooling?

[The Stilt]

Yeah, Trinity and some Richland models are broken.

 

Will fix it in a new version, with couple of new SKUs added.

[mekduk]

Thank you very much for your very prompt reply. Looking forward to the newer version to check the other APUs that I have on hand.

[flanker]

will be working with Kaveri chips too or it must be new software for it?

[mekduk]

Just a wild guess, when the BIOS/UEFI is set to default CPU voltage (auto) on stock clock, would it be an indicator of rough estimate on the base leakage percentage?

[subaruwrc]

Simply awesome effort there !

[The Stilt]

AFI R1.02

 

Changes:

 

- Fixed the issues with Trinity APUs.

- Fixed the faulty PState7 frequency reading.

- Added support for the new Richland APUs / NPUs ("20"-series, K-series).

[The Stilt]

will be working with Kaveri chips too or it must be new software for it?

 

The software will not run on Kaveri due the hardware checks.

The registers on Kaveri and Carrizo are different anyway.

 

AFI is likely to be the last software I make so I cannot promise support for the upcoming parts.

If I happen to be still around I will add support for Indus, Bald Eagle, Steppe Eagle, Mullins and Golden Eagle too.

[mekduk]

You are indeed the legend

 

AMD should just channel a tiny bit of their resources just for this kind of project. Works better than the huge marketing budget that they have spent so far.

[mekduk]

Quick question, to achieve a good result for undervolting/underclocking A6-6400K (target clock is ~3Ghz), what to look for in this AFI tool (i.e.: how to bin for the lower power?)?

 

Here is my current result on my A6-6400K attached.

[The Stilt]

That is a low leak part, which are generally optimal for underclocking.

The required voltage will be slighty higher, however the current draw is much lower.

 

Puts less stress on the VRM too.

 

After the estimated load-line effect, the actual stock voltages would be (roughly):

 

3.9GHz - 1.29875V

3.6GHz - 1.24125V

3.1GHz - 1.15625V

 

Only beyond that it is undervolting (technically).

Those frequencies and voltage should be "factory stable" with load-line close to 0%.

[mekduk]

Thanks so much for your concise and quick reply. How about the other voltages that we can set through UEFI?

 

At the moment not planning to use it for gaming but solely files/music playback.

 

Currently stable at 1.000V at ~3.1Ghz (slightly less, got the board with 99.82Mhz

bus speed). Really interesting to see the upcoming APUs and what the possibilities would be with the smaller process some time early next year.

[flanker]

it will be question...Its TSMC 28nm, hard to say....I heard some spekulation about 28nm TSMC and FX: one of the reasons no more FX yet is technology process by TSMC. In theory 28nm TSMC is more power hungry than current 32nm SOI HK with the same size of die. So maybe some FX with new process (20nm?)???

[The Stilt]

it will be question...Its TSMC 28nm, hard to say....I heard some spekulation about 28nm TSMC and FX: one of the reasons no more FX yet is technology process by TSMC. In theory 28nm TSMC is more power hungry than current 32nm SOI HK with the same size of die. So maybe some FX with new process (20nm?)???

 

Only SoCs and the GPUs are made on TSMC 28nm node.

The APUs are made on GlobalFoundries new 28nm node.

[filmbot]

Work on Llano also? Only seen mention of Trinity and Richland.

[NoMS]

 

Tested 3D on DICE, IGP isn't that great on this one (+/- 1350 3dm03 @ 1.65V).

 

EDIT: Nothing is good on this one...

Edited November 13, 2013 by NoMS

[Calathea]

Exactly the same values on my new 5800k. Seems there still are some issues when using trinity.

[Splave]

should be not bad but not great correct?

[FqqTBawL]

Hi, Sorry to bother. I can not seem to get this working? I keep getting an application error "memory could not be read" as per the attached picture.

 

Asrock A88x Extreme 6+ and A8-6600k windows 7 32

 

Got it working, fresh install windows...

Edited February 1, 2014 by FqqTBawL