What is Richland? AT users vote.

[inf64]

So now that we know Kaveri(28nm) is a 2014 product and Richland took its place, what do you guys think Richland is?

[ShintaiDK]

What if 28nm gives a clock decrease instead from going SOI to Bulk?

But due to the GloFo agreement option 2 is much more pausible. Even tho there are better economics in option 3, had the GloFo agreement not been in place.

Option 1 and 4 is exculded from AMDs own information.

[inf64]

Originally Posted by ShintaiDK

What if 28nm gives a clock decrease instead from going SOI to Bulk? Else I would vote 3. Unless AMD picks 2 due to the GloFo agreement.

Sure that's possible. But what would warrant a 6000 model number if CPU/GPU sides would "suck" when compared to 2012 5000 series?
I think in this case I could redefine the poll with "possible clock bump" for option number 3/4.

edit: Sorry I can't modify the poll,no option that I can see.

[ShintaiDK]

Originally Posted by inf64

Sure that's possible. But what would warrant a 6000 model number if CPU/GPU sides would "suck" when compared to 2012 5000 series?
I think in this case I could redefine the poll with "possible clock bump" for option number 3/4.

edit: Sorry I can't modify the poll,no option that I can see.

You could say GPU wise it would be warranted.

[inf64]

Yeah but they cannot screw things more than they are now . I even have doubts this thing will be 28nm at all(but still voted for 3) . They will make some changes in IMC to bring more memory BW which points out to somewhat better GPU,but CPU side will probably be the same and only thing that can make it better is clock increase to say ~4Ghz. But since Trinity almost always hits 4Ghz in normal desktop stuff users run on it, Richland would have to reach higher on the clock scale to make any difference in common benchmarks. This will be hard to do no matter what node it's being made on.

[ShintaiDK]

Thinking more about it. I am actually positive that Richland is 32nm and with CGN cores.

AMDs own slides already excluded 1 and 4 since it will be CGN cores and Pilediver. And by the colour used for Richland its supposed to be 32nm. Exlcuding option 3 as well. So only number 2 left.

(I actually misclicked for option 3 myself.)

[CHADBOGA]

I think Richland is likely to be one of the most insignificant products of all time.

[NTMBK]

Originally Posted by CHADBOGA

I think Richland is likely to be one of the most insignificant products of all time.

This.

It will probably make as much impact as Brazos 2.0 did.

[inf64]

Originally Posted by ShintaiDK

Thinking more about it. I am actually positive that Richland is 32nm and with CGN cores.

AMDs own slides already excluded 1 and 4 since it will be CGN cores and Pilediver. And by the colour used for Richland its supposed to be 32nm. Exlcuding option 3 as well. So only number 2 left.

(I actually misclicked for option 3 myself.)

Yeah but I don't think changing votes is possible .
I hope we will know more after CES 2013.

[Idontcare]

I chose option 3 "28nm bulk,PD based,GCN GPU core(HSA),clock bump on CPU and GPU side" as a balance between what seems possible (albeit not the most probable) and what I wish/hope AMD is able to do.

I do not see a "PD+" core being at all possible, neither Llano nor Trinity carried "+" versions of the CPU cores at the time and I doubt Richland would either.

Originally Posted by inf64

Sure that's possible. But what would warrant a 6000 model number if CPU/GPU sides would "suck" when compared to 2012 5000 series?
I think in this case I could redefine the poll with "possible clock bump" for option number 3/4.

edit: Sorry I can't modify the poll,no option that I can see.

Originally Posted by inf64

Yeah but I don't think changing votes is possible .
I hope we will know more after CES 2013.

If you want the poll text changed then it can be changed by a mod, just let me know in exact detail what you want changed (specifics in regards to "from -> to").

If you want a new poll that is possible as well. Start a new thread with your new poll exactly as you want it and then let me know. I will merge this thread with the new poll thread, and in doing so I will keep this thread (all the posts) but I will keep the newer thread's poll.

It basically substitutes the existing poll with an entirely new one, causing this thread's poll to disappear as well as the other thread (the new thread you'd make w/ the right poll) to disappear.

[pablo87]

I pick #2

[Abwx]

Picked 2....

It wouldnt be a 2.0 version if only frequencies are the difference ,
while a 28nm version seems a bit early given that the current process
is barely two years old in respect of industrial scale production.

[VirtualLarry]

I hope it's 3, but I picked 2, being a little more realistic.

[IntelUser2000]

I chose #3.

28nm, because previous roadmaps on Kaveri was indicating 28nm. PD, because I agree with IDC's post. GPU is probably not that much different from what was supposed to be in the original Kaveri. It's Trinity's CPU cores + Kaveri GPU on a 28nm process. While clocks might have to go down, 28nm will offer some density increases, keeping Richland at similar die sizes to Trinity despite having enhanced GPU.

[Centauri]

Hoping for 4, voting for 1.

[Haserath]

4, good luck

3, unlikely

2, probable

1, wouldn't surprise me

I'm wavering between 1&2. Kaveri was planned as SR and GCN, but Richland only recently replaced it. Makes me think it's a filler with no time for change over Trinity. Graphics are a bit easier to bolt only Trinity cores though, especially since they have them worked out already... But that's on 28nm bulk, and Trinity(PD) isn't on 28nm bulk...

I choose 1 as most probable.

[parvadomus]

Picked 2. Sadly this is Kaveri's replacement for 2013. It will pack GCN cores and hybrid cross fire Mars XTX based low end GPUs.
And the only reason for AMD to launch this is to counter Haswell IGP performance. Steamroller will be ready and polished for 2014 (or skipped directly to excavator or SR+).

[jpiniero]

No option for Cancelled?

[Centauri]

Why would there be?

[Homeles]

Originally Posted by Centauri

Why would there be?

Gotta start our inane AMD-bashing circlejerk somehow.

[inf64]

Originally Posted by jpiniero

No option for Cancelled?

Richland cancelled? You are not following AMD then I suppose . Richland is filling in for missing Kaveri on 2013 roadmap . If they cancel the replacement ,what would they use as a replacement for a replacement?

Richland is coming for sure, we just don't know what it is exactly hence this little poll. Soon we will know who had hit the closest to the real spec .

[NostaSeronx]

Originally Posted by ShintaiDK

What if 28nm gives a clock decrease instead from going SOI to Bulk?

The clock decrease will come with a much lower voltage with it.

[sm625]

If you know AMD then you are expecting 1 and hoping for maybe 2 lol.

[ViRGE]

Originally Posted by ShintaiDK

Thinking more about it. I am actually positive that Richland is 32nm and with CGN cores.

AMDs own slides already excluded 1 and 4 since it will be CGN cores and Pilediver. And by the colour used for Richland its supposed to be 32nm. Exlcuding option 3 as well. So only number 2 left.

Ditto. That ended up being my exact train of thought.

[Idontcare]

Originally Posted by NostaSeronx

The clock decrease will come with a much lower voltage with it.

The technical issue with going "32nm SOI/HKMG -> 28nm bulk/HKMG" is not that Idrive will remain flat or go down (because it won't, Idrive assuredly will go up with 28nm over 32nm) but the static and dynamic leakage value ought to markedly increase with the 28nm non-SOI process over that which AMD is already relying on with the existing 32nm SOI process.

That much must be true unless the existing SOI-implementation in 32nm is so shoddy, so crappy and so un-optimized, that its removal from the process flow results in nary a difference in static or dynamic leakage.

So we must either assume them to be fools for going to the expense of using SOI at 32nm, or assume them to be fools for bothering to pay for taping out chips for 28nm that don't have SOI, because process-technology wise this should result in little more than trading six of one thing for a half dozen of another.

The primary benefit here from such a node transition "32nm SOI/HKMG -> 28nm bulk/HKMG" is AMD and the cost per die. The wafers will certainly yield more dies/wafer because the 28nm xtors will have higher density (smaller die) and without the SOI tax the wafers will only be slightly more expensive than the existing 32nm ones.

So AMD gets to sell chips which cost them much less per chip, but the chips themselves would not be expected to perform any better (nor worse) than today's chips in terms of clockspeeds and power consumption - but there is the opportunity for IPC improvements of course.

(unless 32nm SOI is just so badly borked that 28nm without SOI does actually turn out to still be just that much better...)