[Overclockers.com]

It's not every day we look at a 'mainstream' part. AMD's soon-to-be-released Lynx platform (Llano architecture for desktop systems) is as mainstream as it gets. Targeted for a total system cost of $500-$600, this platform is geared toward the biggest market segment for PCs today.

While many of our readers wouldn't go for this platform as their main PC, many of us have other uses for Lynx - HTPCs for one. Let's see what AMD has cooked up for us with their latest accelerated processing unit.

... Return to article to continue reading.

[Evilsizer]

looking good from amd!

though this does not bold well for NV, since the onboard IGP is there on the cpu now. doesnt look like NV will be able to release a chipset with IGP for either amd or intel. looks like the only thing they will be able to work on is the tegra line and possibly the ION line. however i feel like nv will drop the ion line in favor for more RD towards the Tegra line.

For dirt 2 i didn't know that high settings was needed, as the way i read it it was only medium. the sun shafts in stalker is what took me off guard, i ran it twice and still got 2.3fps but the others where higher. i couldn't figure out why either, really odd.

[hokiealumnus]

Now we're cooking! See, just needed a little more time than I had before the NDA expired.

[Evilsizer]

wow that is some voltage for 3.6ghz... no way to oc the IGP?

[hokiealumnus]

IGP overclocks with the bus. You can't manually raise IGP unfortunately. Like the CPU, it has a locked multiplier. In this case, it runs 600MHz at stock (100MHz bus), so at 133 bus, the IGP runs 798MHz.

As far as voltage, that's only 0.1V above 1.4V stock. Since it started at 1.4, I figured .1 can't hurt too badly.

[Evilsizer]

o, lol, im to use to intels lower Voltage levels. i guess then that doesnt sound to bad, so with a %20 safety factor above stock. you would/could try up to 1.68v but the question is how well does the cpu respond to voltage. how would it respond to other voltage increase before touching the cpu voltage. since the IGP gets overclocked as well does that mean more cpu voltage to keep the IGP running at the higher speeds or is there a seperate IGP voltage setting.

[hokiealumnus]

No separate IGP voltage on this board anyway. I've hit a wall at 144MHz, then video just stays black. With multi-adjusted CPU OC to 3.6 GHz, I'm testing to see if this is the sweet spot. It would put the IGP at a decent 864MHz.

[Evilsizer]

any chance you can disable IGP and see how well the cpu oc's on its own?

[hokiealumnus]

I could do that without worrying about the IGP within reason. It defaults to a 29x multi. 140 x 29 = 4060 MHz. If that's stable, I can try another GPU.

Having some trouble at 144MHz bus though. It won't make it very long in Prime and won't make it through LinX. The problem seems to be thermal throttling...not so smart thermal throttling. When the CPU gets to 65 degrees in CoreTemp, it raises the multiplier from 25x to 26x, giving an extra ~150MHz, which flunks it every time.

It might not be thermal though, b/c raising CPU NB and CPU PLL does help LinX go longer, so it may just be a matter of some voltage adjustments. Will keep working on those as time allows. The kiddo's up now though, time for some family fun time!

[hokiealumnus]

Turns out the 25x multi was just wonky and I needed to use the 26x multi to take it as far as it would go. I'll post up the LinX screenshot and one benchmark....the rest I'm running and will publish a Llano/Lynx overclocked article when I'm done.





Not to shabby clocks for a $135 CPU on air. 142MHz bus = 852MHz on the IGP. Could push it to 144 bus but the CPU wouldn't go that far without dropping to the bad mojo multi. I'd rather sacrifice fewer GPU MHz for the 142MHz CPU bump.

[bigben6]

so is this is bulldozer series or not, i'v been following casually, this is like ebtry level BD correct?

[hokiealumnus]

Negative, this has no BD parts in it at all. This is still a modified Stars core like those found in Phenom II and Athlon II CPUs. The big deal about Llano is their IGP. There's nothing especially innovative about the CPU part of the die.

[bigben6]

ok thanks, still at that price point.... its tough to beat, and 8 cores + IGP that runs Dx11 for ~150 is a heck of a deal for a HTPC/casual gaming rig

[EarthDog]

AWESOME review! Nice IGP product!


We really need to get pages on reviews though!!

[nookuwn]

Originally Posted by bigben6

ok thanks, still at that price point.... its tough to beat, and 8 cores + IGP that runs Dx11 for ~150 is a heck of a deal for a HTPC/casual gaming rig

Where did you get 8 cores from?

[bigben6]

Originally Posted by nookuwn

Where did you get 8 cores from?

im dyslexic and read that a6 and a8 and just typed without comprehending... again thinking it was bulldozer (which has been corrected) with rumors for 6+ core procs the a6 and a8 seemed like obvious nomenclature for 6 and 8 cores respectively... ignore me....

[trents]

Why is the stock bus 100 mhz instead of 200 like is is for other CPUs?

[TheQuadFather]

Originally Posted by trents

Why is the stock bus 100 mhz instead of 200 like is is for other CPUs?

following intel suit maybe, i dont know, maybe 100 x whatever is easier to work out than 200 x whatever, they got MASSIVELY lazy okay?

[trents]

Well, I was just wondering because it would seem to have a significant negative impact on performance.

[Archer0915]

Originally Posted by trents

Why is the stock bus 100 mhz instead of 200 like is is for other CPUs?

Originally Posted by TheQuadFather

following intel suit maybe, i dont know, maybe 100 x whatever is easier to work out than 200 x whatever, they got MASSIVELY lazy okay?

Originally Posted by trents

Well, I was just wondering because it would seem to have a significant negative impact on performance.

BClock, HTT clock or metronome. Just from analysis the 100 is greater than 200. All this is, is a base for multiplication and that is where the FSB confusion starts.

http://www.techreaction.net/2009/11/...-overclocking/

IDK if that will help but basically the 100MHZ is only a base multiplier on newer processors it is not the front side buss. With HTT and QPI the end result is the front side buss if you will. Where the difference occurs is the interactions that are based off of the 100 or 200 where multipliers differ. With memory base this may not hold true but to all other busses be they divided or multiples this applies.