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

The number of CUs by themselves, either in a console or a discrete part are not that important, IMO. I just hope they beef up the caches and make them flexible.
 
If according to google, TSMC is claiming to get 20nm going in the second half of 012, why would AMD be planning a 2013 Sea Islands GPU on 28nm?
 
if you look at your first link, it doesn't even suggest they'll have an a-15 out on 20nm this year. The first A-15s are going to be 28nm. They've been doing tapeouts and testing for 20nm, but that doesn't really mean it's going to be ready for production of complex product like a gpu this year.
 
what are they going to be producing on 20nm in the later half of 2012? SRAM?

There's that, and…

If according to google, TSMC is claiming to get 20nm going in the second half of 012, why would AMD be planning a 2013 Sea Islands GPU on 28nm?

…these roadmaps for the financial crowd are usually very coarse, where 2013 means it's going to be the main 2013 graphics family, but the first chips could very well surface in Q4 2012.
 
If according to google, TSMC is claiming to get 20nm going in the second half of 012, why would AMD be planning a 2013 Sea Islands GPU on 28nm?

As they did in the past 3 years, they launch Sea Island in Q4-2012/Q1-2013, and I'm pretty sure that 20nm won't be ready until end of 2013. The successor of Sea Island, Q4-2013/Q1-2014 will use the 20nm process.

What kind of improvements can we expect from GNC 2.0?
 
Not sure why not. Tahiti was just a month and a half over 1 year from Cayman. And that considering it's a complete revamp of the architechture. Cayman was about a year after Cypress. Cypress about a year after Rv770 (er wait or did 4890 make cypress longer than that, don't remember and don't feel like looking it up).

So approximately one year for Sea Islands with similar arch shouldn't be odd.

AMD has been executing far better than Nvidia ever since GTX 280/Radeon 4870.

Or at least with far shorter time between products (except GTX 480->580).

Perhaps it's due to working with smaller dies. Nvidia themselves did say it was f****** hard to get a big die GPU out. :)

Regards,
SB
 
Not sure why not. Tahiti was just a month and a half over 1 year from Cayman. And that considering it's a complete revamp of the architechture. Cayman was about a year after Cypress. Cypress about a year after Rv770 (er wait or did 4890 make cypress longer than that, don't remember and don't feel like looking it up).

So approximately one year for Sea Islands with similar arch shouldn't be odd.

AMD has been executing far better than Nvidia ever since GTX 280/Radeon 4870.

Or at least with far shorter time between products (except GTX 480->580).

Perhaps it's due to working with smaller dies. Nvidia themselves did say it was f****** hard to get a big die GPU out. :)

Regards,
SB

And Cayman had to be backported from 32nm to 40nm, which certainly must have delayed it at least a little. Sea Islands will be on 28nm, so I don't see why we should have to wait significantly more than a year; it might even be less.
 
AMD did state at the analyst day, that 28 nm was high k metal gate across the board.
Sure. But SOI HKMG or bulk HKMG for the Trinity successor?

As GF appears to be doing 28nm SOI, they could also use that process.
 
My guess is they will stay with SOI for that class of product. It would really really have to be worth while for them to make a move, im guessing its the same on the GPU front as well.
 
quoting Thomas Ryan:

Back yet again to the topic of 28nm, an analyst tried to tease more information from the session, “I am going to try and re-ask the question on GPU 28nm to see if it’s High-k or SiON, because TSMC supplies both.” He followed that question up with an even more important one, “Your roadmap this morning laid out 28nm, in 2013 across, almost all of your products. Will you lean toward a more High-k [process], or more SiON [process], or a mix?” The answer to these two questions was both short and firm from Mr. Akrout, “Since you asked twice we are going to get you an answer. All of them are in the High-k metal gate. Across the board.”

http://semiaccurate.com/2012/02/07/fabless-works-for-amd/
 
Last edited by a moderator:
Back
Top