Awesome Llano Overclocking Tool

[HTC]

When you update it , tell me whats fixed and does it work properly

In case you're wondering, AGESA is:

"...The AMD Generic Encapsulated Software Architecture (AGESA) is a
library of validated processor procedures designed to aid customers with
quick adoption of AMD technology into their products. AMI’s drop-in
support of AGESA allows minimal time-to-market for APTIO® customers to
integrate AMD Processor Solutions into their products. AMI’s integration
in APTIO® indicates that AGESA™ code is versatile enough that it can be
dropped in totally new architecture (UEFI) with minimal build changes...."

Source

This new BIOS may indeed resolve this issue: i'll probably update it Sunday.

EDIT

Forgot the small detail where it says the only thing my board's BIOS is updating is the AGESA code, which is why i posted the 1st part of this post ...

 

[HTC]

When you update it , tell me whats fixed and does it work properly

This new BIOS may indeed resolve this issue: i'll probably update it Sunday.

Just updated it and tested: seems to NOT have worked: CPU-Z shows "overclock" but IBT doesn't go faster then stock, like with previous BIOS

Whatever they changed in the AGESA code, i have yet no notice anything different with this new BIOS.

In attach are a couple of pics from my BIOS: the only OC i'm doing right now is with the RAM, and even that is just a tad overclock (timings, not speed).

 

[InternalSys]

Just updated it and tested: seems to NOT have worked: CPU-Z shows "overclock" but IBT doesn't go faster then stock, like with previous BIOS

Whatever they changed in the AGESA code, i have yet no notice anything different with this new BIOS.

In attach are a couple of pics from my BIOS: the only OC i'm doing right now is with the RAM, and even that is just a tad overclock (timings, not speed).

Lol, my BIOS is so different from your BIOS, Asrock made their bios like a Windows 7 APP , its not bios like - bios anymore. And i managed to overclock my ram up to 2400 mhz, from 1333, but iam not sure if it is overclocked for real , is there any bench to test it, so i can see if theres any improvment, cos when i tested it on my windows 7 test i didnt show any improvement hehe

 

[HTC]

Lol, my BIOS is so different from your BIOS, Asrock made their bios like a Windows 7 APP , its not bios like - bios anymore. And i managed to overclock my ram up to 2400 mhz, from 1333, but iam not sure if it is overclocked for real , is there any bench to test it, so i can see if theres any improvment, cos when i tested it on my windows 7 test i didnt show any improvement hehe

Because of my hardware RAID card, i can't have a board with UEFI BIOS only: incompatible

This severely limited my choices for an APU board. This is the 1st non ASUS board since 1996.


As for your question, try AIDA64 or MaxMem: both of these should show if your RAM is actually OCed or not.

 

[cadaveca]

Just FYI, Llano desktop chips do not actually change multi upwards from stock. It will appear that it has changed, but it doesn't have a real effect. Software tools will not help either. Every board I have tested works this way, and I have checked with other reviewers taht I talk to, and they find the same thing.


For underclocking, sure works great, but not for OC. You must increase HTT for OC.

 

[Mathragh]

Just FYI, Llano desktop chips do not actually change multi upwards from stock. It will appear that it has changed, but it doesn't have a real effect. Software tools will not help either. Every board I have tested works this way, and I have checked with other reviewers taht I talk to, and they find the same thing.


For underclocking, sure works great, but not for OC. You must increase HTT for OC.

I wonder why laptop APU's and Yo_Wattup's chip doest seem to be locked then

 

[cadaveca]

They are not "locked". The multi will appear to change in CPU-Z, but performance will not change at all.

I recently was working on a review of a Llano board, and the same applies. Doing some research led to other people having hte same problem...on some obards, you can just change multi in BIOS, it appears to work, but again, no performance change.

Of course, this seems confined to the A4, A6, and A8 chips, the other APUs, I am not sure of.

 

[Mathragh]

They are not "locked". The multi will appear to change in CPU-Z, but performance will not change at all.

I recently was working on a review of a Llano board, and the same applies. Doing some research led to other people having hte same problem...on some obards, you can just change multi in BIOS, it appears to work, but again, no performance change.

Okay, but the doesnt change the fact that i do get a measurable performance increase on my mobile Llano APU, and so did Mussels. In fact, I was able to get a 30% performance increase while not substantially increasing power draw, since i could lower the voltage a whole lot, even at a higher frequency. The screenies for this are at the start of the thread, and some further on.

Edit: So i wonder what makes those chips different from the ones that doesnt seem to be able to OC with this tool. Perhaps its actually the motherboard? I dont know.

 

[cadaveca]

Mobile, sure, desktop, no. I am not 100% sure why there is a difference.

Performance increase without power draw increase kinda points at something else, IMHO. You cannot magically get added performance without an increase in power consumption. When transistors switch faster, they need more juice to do so.

I've seen myself results from 8 different boards, and I'm talked to other board reviewers at other sites about at least another 5 boards.


All desktop though.

 

[Mathragh]

Hmm, well, thanks anyway for sharing your experience with desktop APU's Will provide some clarity for desktop Llano users.

EDIT:
The power use did increase when i upped the frequency. In fact, I didnt dare up the frequency as high as I have it now, since the laptop would get too hot. With the voltage decrease at the same time as the higher mult, the cooling could sustain the higher clocks.

 

[cadaveca]

Hmm, well, thanks anyway for sharing your experience with desktop APU's Will provide some clarity for desktop Llano users.

EDIT:
The power use did increase when i upped the frequency. In fact, I didnt dare up the frequency as high as I have it now, since the laptop would get too hot. With the voltage decrease at the same time as the higher mult, the cooling could sustain the higher clocks.

Ah, I understand.

And yeah, you know, I will be mentioning this in my next A75 board review. I've already compiled results, with benchmarks, screenshots, and such, but I do nto think I will include all of that in the review, as I'm reviewing the board, not the APU.

Of course, the A8 3870 BE or whatever they call it..."K", maybe? does increase performance when multi is adjusted(using the same boards). It's a very weird situation, this multi thing on the "locked" desktop chips.

 

[Mathragh]

Ah, I understand.

And yeah, you know, I will be mentioning this in my next A75 board review. I've already compiled results, with benchmarks, screenshots, and such, but I do nto think I will include all of that in the review, as I'm reviewing the board, not the APU.

Of course, the A8 3870 BE or whatever they call it..."K", maybe? does increase performance when multi is adjusted(using the same boards). It's a very weird situation, this multi thing on the "locked" desktop chips.

Indeed it is. I'm just happy I found out about this tool, as it does loads for me. I'm still really puzzled though why these(or atleast mine) chips has that much headroom in both frequency and voltage. It can consequently do with 0,2V less then what it is put at at stock. You'd think that they would want to make these chips as energy efficient as possible.

Also, i'm still not sure wether this actually works on ALL mobile chips. I think about 4 people on this forum with mobile Llano APU's have already succesfully used this tool to OC their mobile chips, with probably more not posting anything but still being succesfull.

 

[HTC]

I wonder why laptop APU's and Yo_Wattup's chip doest seem to be locked then

Maybe he has a "K" model?

If not, i have no idea

 

[InternalSys]

Will be good, if anyone have an AMD forums account, to go there and make a TOPIC about it , and if someone from AMD explains, it would be great

 

[DaC]

Indeed it is. I'm just happy I found out about this tool, as it does loads for me. I'm still really puzzled though why these(or atleast mine) chips has that much headroom in both frequency and voltage. It can consequently do with 0,2V less then what it is put at at stock. You'd think that they would want to make these chips as energy efficient as possible.

Well, I don't know about the process of making this kind of stuff, but it's all about economics and engineering...

They control a small batch only and by the result of this batch, using statistics, they know more or less how this batch of chips will perform and set their specifications....
Of course there will always be some that work well better than the specifications (this is way overclockers love so much to search the "god sent chip") and others will just run very close to the specification set.
The important thing is to ensure a very low failure rate and because of that all of the chips gets the specification of maybe the worst performer or close to it.

That's why some gpus get 7970 labels while others 7950 label although being the same gpu... sometimes like phenom II the batches are so good that they have to put higher grade chips to lower specifications just to comply with the market strategy.

Checking each individual chip would lead to enormous costs and would make production cost just crazy...

So, people shouldn't blame x or y company for being a poor quality company just because they bought one defective product from them. It happens, the company knows that it will and it should be ready to deal with it... and if it isn't ready to well support their consumer, then it can be called a bad company. Or sometimes failure rate are just so high that there isn't any excuse.... hp notebooks anyone ?

 

[Yo_Wattup]

So I FINALLY got some faster memory (sig), will be benching soon.

 

[Yo_Wattup]

Okay, more strangeness coming from llano, passmark score has decreased to 2700 from 3200.

But 3Dmark vantage has increased to 8300 compared to 7200 stock freqs and 8000 with old ram.

I am inclined to believe passmark is telling lies to me and that I should go off 3Dmark.

 

[cadaveca]

Check benchmarks I use in my memory reviews for benches that wuill show differences from RAM changes. Personally i like Spi32M for that.

 

[DaC]

I don't know if this has happened to you folks using the mobile Llano like me, but I realized that I had to up the otherwise stable vcore after plugging USB devices.
First I plugged a USB mouse on my notebook...... no problem, all voltages remained the same on my overclock result post: http://www.techpowerup.com/forums/showpost.php?p=2566684&postcount=154 although I had already upped my P0 state 3 notches after this post.

Then I plugged my USB soundcard and a hdmi cable to hook it on my monitor.... the system begun to reboot with normal internet use.... youtube and stuff....... so I upped a notch all my P states (probably I wouldn't have to mess with them all, but I wasn't just with the will to test it)..... after that, everything went just smooth...

Today I plugged one more device, a USB (outlet powered) 3.5" hard drive.... the system begun to reboot again while watching some movies on VLC..... So I uppped all p states one more notch and everything seems fine....

It seems to me that the amount of usb ports being used actually interfere with the mobile Llano overclock capacity in the way that you have to use more vcore on it than with less usb ports (and hdmi) being used.....

What do you guys think ?

 

[Mussels]

I don't know if this has happened to you folks using the mobile Llano like me, but I realized that I had to up the otherwise stable vcore after plugging USB devices.
First I plugged a USB mouse on my notebook...... no problem, all voltages remained the same on my overclock result post: http://www.techpowerup.com/forums/showpost.php?p=2566684&postcount=154 although I had already upped my P0 state 3 notches after this post.

Then I plugged my USB soundcard and a hdmi cable to hook it on my monitor.... the system begun to reboot with normal internet use.... youtube and stuff....... so I upped a notch all my P states (probably I wouldn't have to mess with them all, but I wasn't just with the will to test it)..... after that, everything went just smooth...

Today I plugged one more device, a USB (outlet powered) 3.5" hard drive.... the system begun to reboot again while watching some movies on VLC..... So I uppped all p states one more notch and everything seems fine....

It seems to me that the amount of usb ports being used actually interfere with the mobile Llano overclock capacity in the way that you have to use more vcore on it than with less usb ports (and hdmi) being used.....

What do you guys think ?

i'm guessing you get vdroop with power hungry devices. i ran into something similar with mine, took everything up two notches and its stable since then (but apart from my first test, i've always had a USB mouse and keyboard connected)

 

[DaC]

i'm guessing you get vdroop with power hungry devices. i ran into something similar with mine, took everything up two notches and its stable since then (but apart from my first test, i've always had a USB mouse and keyboard connected)

Interesting... so it might be pretty normal with it then...

 

[Mussels]

Interesting... so it might be pretty normal with it then...

a laptop having a weak PSU would not surprise me in the slightest, especially since they had the stock voltages set so high compared to what they really need. looks like they knew this happened, and upped the voltages to prepare for it.

 

[Mathragh]

I didnt have such issues, but perhaps I just got lucky, the psu this laptop uses is quite big and looks fairly heavy duty(I know that shouldn't mean anything, lol).
But anyways, good to know in case others encounter the same problem

 

[DaC]

a laptop having a weak PSU would not surprise me in the slightest, especially since they had the stock voltages set so high compared to what they really need. looks like they knew this happened, and upped the voltages to prepare for it.

Yes, probably. To the point that my P0 state is in the same voltage as the stock, even thought I'm using 3.0ghz instead of the 2.5 stock.
Although I had to raise my P1 state a few notches, it's still a lot lower than stock one and with a 500mhz bump.
But yeah, makers do give a nice margin of safety when they set cpus, which is a good thing btw.

I didnt have such issues, but perhaps I just got lucky, the psu this laptop uses is quite big and looks fairly heavy duty(I know that shouldn't mean anything, lol).
But anyways, good to know in case others encounter the same problem

Well mine is rated at 19v 3.16A... it's fairly small and light... what's the output from yours ?

 

[Mussels]

Yes, probably. To the point that my P0 state is in the same voltage as the stock, even thought I'm using 3.0ghz instead of the 2.5 stock.
Although I had to raise my P1 state a few notches, it's still a lot lower than stock one and with a 500mhz bump.
But yeah, makers do give a nice margin of safety when they set cpus, which is a good thing btw.


Well mine is rated at 19v 3.16A... it's fairly small and light... what's the output from yours ?

i wasnt talking about the brick, i meant internal to the laptop itself - its got to be split into 5V for the USB somewhere, for example.