Guess what product AMD is revealing!

[Abwx]

http://wccftech.com/amd-power-mulli...powered-puma-x86-gcn-cores-launches-tomorrow/

...

 

[rtsurfer]

Slides look legit.

Now waiting for the reviews to see if they can hit the claimed performance/power targets.

 

[monstercameron]

Slides look legit.

Now waiting for the reviews to see if they can hit the claimed performance/power targets.

well be prepared to wait...kabini took months to come to market.

 

[Abwx]

Thoses numbers point to GF 28nm as being quite better than TSMCs same node unless it was fabbed on the latter s eventualy improved process as it s about the same perf/watt improvement one would get with an averagely scaling full node transition.

 

[Phynaz]

TSP? Seriously?

 

[ultimatebob]

Slides look legit.

Now waiting for the reviews to see if they can hit the claimed performance/power targets.

I'd be more concerned about the actual shipping price to OEM's, as Intel will likely try to match it with a competing product.

 

[erunion]

The TDP is 45% lower but the actual power consumption is only 20% lower. 8w vs 4.5w tdp.
2x performance per watt did not happen, as expected.

 

[monstercameron]

The TDP is 45% lower but the actual power consumption is only 20% lower. 8w vs 4.5w tdp.
2x performance per watt did not happen, as expected.

I am sure that claim was scenario and sku dependent.

 

[Abwx]

The TDP is 45% lower but the actual power consumption is only 20% lower. 8w vs 4.5w tdp.
2x performance per watt did not happen, as expected.

I guess that you re not in good terms with %age and numbers..

 

[Shivansps]

One on the AMD slides says 20% power consumption.

So let me see if i got this right, 40% lower TDP, 20% less power, 10% more IGP perf, probably because of the 1866 mem.

Dont tell me thats all, you can ajust TDP values and also get less power in the same way Intel does, i think they are doing the same thing.

 

[piesquared]

Yep, 2x perf/watt like expected. Extremely impressive especially on the same node, AMD pulled off some nice wizardry here! OEM's must be lining up with new designs, it's the best product available for quite some time.

 

[Abwx]

One on the AMD slides says 20% power consumption.

20% of what?..Of the full APU ? Full system ?...

 

[erunion]

I guess that you re not in good terms with %age and numbers..

You guessed wrong.

4.5 is ~55% of 8. Thus TDP has decreased 45%. But the AMD slide says power has decreased up to 20%.

 

[Vesku]

Well if the leaked slides are accurate that A6-6310 is 1923 PCMark 8 it's way off the preview published number of 2312 for 15W TDP Beema. Unless they have a better binned 15W Beema beyond the 6310 then they really missed on performance aspect.

 

[HurleyBird]

Why Puma+ cores? Did they skip the regular, non + Puma cores or something?

 

[Abwx]

You guessed wrong.

4.5 is ~55% of 8. Thus TDP has decreased 45%. But the AMD slide says power has decreased up to 20%.

The 20% are for Beema , the 15W chip, in an unknown situation since they are related to power draws varying from 0.62 to 1.92W wich are obviously numbers that have nothing to do with this chip nominal TDP, i guess that you were prompt to assume that it was for Mullins/Temash even if the slide say Kabin/Beema explicitely...

 

[Vesku]

Temash/Mullins are listing SDP not TDP.

 

[Phynaz]

Well if the leaked slides are accurate that A6-6310 is 1923 PCMark 8 it's way off the preview published number of 2312 for 15W TDP Beema. Unless they have a better binned 15W Beema beyond the 6310 then they really missed on performance aspect.

AMD over promising shouldn't surprise anyone.

 

[Abwx]

Well if the leaked slides are accurate that A6-6310 is 1923 PCMark 8 it's way off the preview published number of 2312 for 15W TDP Beema. Unless they have a better binned 15W Beema beyond the 6310 then they really missed on performance aspect.

The 4.5W chip scores 1591 but still if the first slide is accurate there will be a nice CPU perfs improvement since they claim over 50% higher frequency at nearly half the TDP, i m most interested in this number as it s impossible that they got such an improvement using the process that was used for the first Kabini iteration.

 

[Shivansps]

20% of what?..Of the full APU ? Full system ?...

Only AMD knowns that.

 

[Abwx]

Only AMD knowns that.

That s why i didnt pay much attention to this slide, it s better to wait for the full coverages but from the single channel thingy we can now be almost sure that theses chips will be used in the AM1 plateform in a year or so.

 

[blackened23]

Yep, 2x perf/watt like expected. Extremely impressive especially on the same node, AMD pulled off some nice wizardry here! OEM's must be lining up with new designs, it's the best product available for quite some time.

Nowhere do they indicate 2x performance per watt. It is interesting how they skirt around performance per watt with stuff as ridiculous as "productivity per watt". LOL. Really. That's certainly a new one. But. You won't find performance per watt anywhere in those slides. Just skewed stuff like "graphics" performance per watt (really??) and productivity per watt. Since AMD is so heavily skirting around the industry standard metric for mobile, you gotta wonder how actual devices will pan out. I'd say that since they're skirting around performance per watt so intentionally in their slides, that will probably show up in the chip. And not in a good way.

 

[monstercameron]

That s why i didnt pay much attention to this slide, it s better to wait for the full coverages but from the single channel thingy we can now be almost sure that theses chips will be used in the AM1 plateform in a year or so.

Hopefully they blow through their kabini stock so we get thin ~november.

 

[Abwx]

Hopefully they blow through their kabini stock so we get thin ~november.

Not sure that it could be as early as in 6 months, Beema/Mullins will completely overtake the mobile segment as the OEMs will switch instantly given the apparently much better perfs, hence all the Kabini chips will be thrown in the DT, HTPC and other such utilities or even in entry level mobile offerings but it will take quite a time to get rid of current inventories.

 

[piesquared]

Nowhere do they indicate 2x performance per watt. It is interesting how they skirt around performance per watt with stuff as ridiculous as "productivity per watt". LOL. Really. That's certainly a new one. But. You won't find performance per watt anywhere in those slides. Just skewed stuff like "graphics" performance per watt (really??) and productivity per watt. Since AMD is so heavily skirting around the industry standard metric for mobile, you gotta wonder how actual devices will pan out. I'd say that since they're skirting around performance per watt so intentionally in their slides, that will probably show up in the chip. And not in a good way.

They sure do, and it's a completely valid statement. This will crush bay trail.

So 2x perf/watt in graphics and productivity.