AMD: Sea Islands R1100 (8*** series) Speculation/ Rumour Thread

[lanek]

The problem is that smaller manufacturing processes take way longer to be ready (and especially before they get cheaper to use than the older ones) then before. Even if TSMC could keep their roadmap, they expect 2 years between 16nm and 10nm. And when you thought going from 40nm to 28nm was bad, well I guess their are big chances, that every new step will be worse.

I dont trust it... Samsung is allready on the final phase of the 12nm and 14nm standardized process ( only for low lower, low performance chips, and for Nand etc )) .. But Samsung use a derivated variant for the 22 - 18nm and is ready for production, include HKMG variant for High Performance.. the first process is a mix on 16nm with 22nm, its a bit of a strange thing, using tridef design and different transistor size for obtain the same result.. but they have prove they have allready push it under the 14nm...

Ofc process from GF, Samsung ( who is working with GF ) and TSMC are different..

Next Generation NVIDIA and AMD GPUs Reportedly Delayed to Q4 2013

This is news from last week.

AMD Says HD 7000 Series To Remain Primary Focus – No New GPU Launch in 2013l

I will not trust anything concerning AMD and obviously Nvidia who come from this site... They have maybe the right information, but they just pump on Asian sites like Redquasar, Pci etc what is writtten, and if many chinese site have got the right write by tbe past because they have more liberty from Nda, we have see too many false information there too . So before we got a triple confimration, i will be really cautionous on thoses infos...

 

[UniversalTruth]

And maybe it's good to include this in the discussion:

GlobalFoundries and Samsung will have new process technologies on par with IBM

http://www.eetimes.com/electronics-news/4406408/GloFo-Samsung-race-to-14-nm-as-IBM-waits-for-EUV

SANTA CLARA, Calif. – Globalfoundries and Samsung are in a dead heat to get their first 14 nm production wafers out before the end of the year, aiming to beat rival Taiwan Semiconductor Manufacturing Co. by as much as a year. Meanwhile, an IBM building in New York sits empty, waiting for an extreme ultraviolet (EUV) lithography machine to light the way to the industry’s longer-term future.

Perhaps it's a good idea for AMD to abandon TSMC altogether anf their infinite problems.

 

[Kaotik]

Next Generation NVIDIA and AMD GPUs Reportedly Delayed to Q4 2013

This is news from last week.

Yes, on one hand you have a news piece by one site claiming that "according to our 'sources' both nV & AMD delay to Q4", on the other hand you have AMD stating to their investors that they're bringing new graphics during H1 - which hand do you think weights more?

AMD Says HD 7000 Series To Remain Primary Focus – No New GPU Launch in 2013l

It's the same "news" piece, with added comment from Twitter.
Now, take a moment to think about it - You're PR of the a company which aims to do as much profit (ie sell as many cards always) as possible - would you ever, ever tell that yeah, our gfx is coming then and then publicly and jeopardize current generation products sales? No, you wouldn't, not even if you were to launch new products tomorrow.

 

[kalelovil]

http://www.4gamer.net/games/135/G013536/20130208088/screenshot.html?num=003

Have nothing new to launch means "stable" in AMD marketing speech.

Even if we were to ignore next generation cards, Mars (384 shader GCN1) has already been designed and will be in mobile products very shortly.
That the slide states that Mars won't come to the desktop in any form during 2013 makes it highly incredulous to me.

 

[Ninjaprime]

And maybe it's good to include this in the discussion:

GlobalFoundries and Samsung will have new process technologies on par with IBM

http://www.eetimes.com/electronics-news/4406408/GloFo-Samsung-race-to-14-nm-as-IBM-waits-for-EUV


Perhaps it's a good idea for AMD to abandon TSMC altogether anf their infinite problems.

So... they don't even have 20nm up and running yet but somehow in less than 10 months they will skip 20nm and have 14nm done to the point of production? LOL.

 

[iMacmatician]

That the slide states that Mars won't come to the desktop in any form during 2013 makes it highly incredulous to me.

Mars may not show up in the retail desktop market, but the desktop HD 8570 OEM and HD 8670 OEM both use the Mars chip.

 

[UniversalTruth]

So... they don't even have 20nm up and running yet but somehow in less than 10 months they will skip 20nm and have 14nm done to the point of production? LOL.

Yes, the point being is that they have to in order not to lag too much behind Intel.
If Intel is able to deliver, then the others are able to deliver the same too. The key word is money, not some magic/ wizard...

Overall, GloFo did a good job in their keynote. They essentially came clean on the early 32/28nm problems that were nagging their reputation and laid out a clear path forward. The pulling in of 14nm may leave some people nervous, but given how much of the process is carried over from 20nm, as long as 20nm is OK, 14′s pull in should not be a big worry. The only problem we have with the roadmap as it stands is that 14XM really seems like a cheat, and that image is unlikely to be shaken until the chips that use it come out. At that time, the results will speak for themselves, until then, whispers will rule.S|A

Global Foundries lays out their FinFET plans
Common Platform 2013: Welcome to the new 14XM and 10XM processes

 

[Ninjaprime]

Yes, the point being is that they have to in order not to lag too much behind Intel.

Do they have a wizard?

 

[AlexV]

The key word is money, not some magic/ wizard...

If I gave you Intel's money right now, you'd manage to catch up to them roughly one day after never. In effect, if GF got the same financial resources allocate to it, it would take a very long time to catch up, if ever. A skilled workforce and a pipeline that has been optimized year after year is not something you buy with a big wad of cash dropped on a counter, unlike what you might believe from Donald Trump / 50 Cent.

 

[sheepdogexpress]

Really, compete with Intel?

Yes, the point being is that they have to in order not to lag too much behind Intel.
If Intel is able to deliver, then the others are able to deliver the same too. The key word is money, not some magic/ wizard...

Overall, GloFo did a good job in their keynote. They essentially came clean on the early 32/28nm problems that were nagging their reputation and laid out a clear path forward. The pulling in of 14nm may leave some people nervous, but given how much of the process is carried over from 20nm, as long as 20nm is OK, 14′s pull in should not be a big worry. The only problem we have with the roadmap as it stands is that 14XM really seems like a cheat, and that image is unlikely to be shaken until the chips that use it come out. At that time, the results will speak for themselves, until then, whispers will rule.S|A

Global Foundries lays out their FinFET plans
Common Platform 2013: Welcome to the new 14XM and 10XM processes

Intel isn't globalfoundaries competition. Unless Intel starts making chips for other companies, they are not stepping particularly on GF toes. Its TSMC and Samsung that GF wants to compete with. Global foundaries doesn't give a damn about AMD. Look how much financial help they have been to AMD in terms of acts of good will. They want to be an independent fab business and get much bigger contracts than AMD can provide.

i.e
Does it really make any sense to spends over 10 billion dollars on Fab because AMD needs its to compete with Intel? Heck no. AMD could never becomes a big enough customer for GF to make such a capital expenditure worth while.

 

[UniversalTruth]

If I gave you Intel's money right now

If you give me so much money I would simply go and buy Intel with everything they have inside.
Or, in the worse case, I will give them enough money and copy all they have. If you say that everything they have is unique and no one else is capable to achieve it...

Intel isn't globalfoundaries competition. Unless Intel starts making chips for other companies, they are not stepping particularly on GF toes. Its TSMC and Samsung that GF wants to compete with. Global foundaries doesn't give a damn about AMD.

I am sure GlobalFoundries do give a damn about every possible customer incliding AMD, NV, Apple, etc.
Intel do produce chips which they want to put inside phones, which would mean their technology is a direct competitor to all fabs producing ARM.

 

[AnarchX]

The roadmap is confirmed: https://twitter.com/AMDRadeon/status/300010683679006720

Maybe we see some GCN2 GPUs for mobile plattforms in Q2 (Neptune, Sun)?

 

[DavidGraham]

AMD IS screwed then ... customers too ..

 

[gongo]

Roadmap ends at Q3 2013? ...so the rumor Q4 new gpu launch aye? AMD hands are tied with the final Orbis and Durango launch from the looks of it...
I dont see how AMD and customers are screwed...i think the only game this year needing a faster single GPU is Crysis 3.... I am glad my 7970 is "useful" for another year! Looking forward to 'dat new memory optimized Catalyst driver!

Granted...anyone remember the leaf blower FX5800...that was the result of Xbox...partially. The curse of the consoles lives on?

 

[Kaotik]

Granted...anyone remember the leaf blower FX5800...that was the result of Xbox...partially. The curse of the consoles lives on?

FX-series failure has nothing to do with XBox

 

[Alexko]

The roadmap is confirmed: https://twitter.com/AMDRadeon/status/300010683679006720

Maybe we see some GCN2 GPUs for mobile plattforms in Q2 (Neptune, Sun)?

That's not a roadmap, it's empty!

 

[DuckThor Evil]

After the news came out about the OEM rebrandings, it's not really surprising that the actual new stuff is not coming soon.

 

[UniversalTruth]

Is it possible that AMD doesn't have anymore technical and financial capability of launching anything new in future?

 

[Homeles]

I dont see how AMD and customers are screwed...i think the only game this year needing a faster single GPU is Crysis 3.... I am glad my 7970 is "useful" for another year

Slowed technological progress is never a good thing. Stop being selfish.

 

[jimbo75]

What is AMD's incentive to release new cards? They are already faster and cheaper than the competition but still selling worse. Should they now release even faster cards that still sell worse, except with bigger dies so it just costs them more?

All we heard after the 7-series releases was "best to wait on Nvidia cards before buying". Well, let Nvidia release their next series and AMD can counter it this time. I find it unlikely that Nvidia is going to refresh top to bottom in the next 6 months so it's not like the status quo is going to be majorly affected.

If the 8-series was only a 15%-20% increase they'd have got panned for it and the press would be telling us to...wait on Nvidia's cards. Where is the incentive to release them now? Might as well wait another 6 months and get the extra performance out of it after seeing what Nvidia's offerings are.

I lost the link but many months ago somebody out of AMD was saying that they wouldn't necessarily be looking at beating Nvidia in time-to-market in future. If your current strategy isn't working then you need to change it.