What is AMD Up To?

[BlockheadBrown]

I'm taking this from another site, but this pretty much sums up what I've been thinking. AMD have "design" (price) wins on all the next gen consoles. Tahiti & Pitcairn are out, but don't have any real new compelling tech (not bashing, btw). Smaller, cooler, faster - good. GCN - okay. Past that, what is AMD up to? I remember when ATI and Nvidia were teasing us with tech demos of what was to come. Now we just get these snippets on release day. That's fine. It leaves me wondering if they have decided to let the graphics division coast (no pun intended) through the next console generation?

Nvidia stated that GPUs are no longer their primary focus. Is AMD going the same route? I'm not asking if this makes business sense. I'm sure it does. AMD's recent acquisition was a bold smart move. As far as graphics are concerned though, have we reached a point where AMD and Nvidia aren't going to be fighting against each other as they once did (mainly ATI & Nvidia). Have the powers that be reached a point of complacency compared to previous graphics technology advancements so that they can focus on newer greener pastures?

Anyway, here's the link: http://www.fudzilla.com/home/item/26606-godfrey-cheng-leaves-amd-after-12-years

...With the last few key people including Rick Bergman, John Bruno - designed Trinity, Eric Demers and Carrell Killebrew, some of the key people for GPUs we don't know if AMD has a plan to remain Nvidia's competitor in graphics business. Most AMD execs who left the company over the past couple of years were all ex-ATI people.

It is evident that CEO Rory Read is forging a new company out of the curios mix of ex-ATI and AMD executives and engineers...

Just pondering here....

 

[mak360]

I for one, welcome this change. Cheng was very good at his job no doubt but I think he needed a break or a new challenge according to S/A


This is a cut throat market and 12 years of it can take its toll on any person.


http://semiaccurate.com/2012/04/02/godfrey-cheng-leaves-amd/

 

[Stuka87]

What recent acquisition has AMD made?

 

[BlockheadBrown]

What recent acquisition has AMD made?

http://sites.amd.com/us/promo/stave/Pages/seamicro.aspx

 

[Stuka87]

ooOOOooo right, forgot about them

 

[Olikan]

interposers for the win!

bandwidth won't be a problem for this child
http://semiaccurate.com/2011/10/27/amd-far-future-prototype-gpu-pictured/

 

[Lonyo]

Obviously they are going to stay in the "GPU" market, but the "GPU" market might not stay as it is now.
The future is heterogenous multi-core processors, and GPUs make up half of that at the moment, and they are turning to be used in more general purpose situations. AMD is pretty reliant on their GPUs at the moment to give them a leg up over Intel in terms of platform solutions.

They have also made various roadmaps fairly clear to indicate where they are going.
High performance GPUs might not make the headlines, but they are still going to make GPUs for at least a few years yet.

 

[RussianSensation]

As far as graphics are concerned though, have we reached a point where AMD and Nvidia aren't going to be fighting against each other as they once did (mainly ATI & Nvidia).

I am not sure I agree. NV is firing on all cylinders in GPUs. Kepler has more design wins in laptops than Fermi, their GTX680 just recaptured the performance crown in absolute performance, features, and key metrics AMD used to hold (Eyefinity no longer an advantage, performance/mm^2, performance/watt, performance/die). NV is going to launch full blown GK110 at some point. They are also working on launching Maxwell in 2014. It might seem like NV has slowed down but part of that stems from 28nm capacity issues and slower transition betwen new node shrinks. AMD is also not standing still. Sea Islands (GCN 2.0) is scheduled for Q1 2013.

It seems most ex-ATi/AMD employees are being rewarded with more lucrative job offers elsewhere. Also, it could be a lot of them are not aligned with the vision of the new CEO. Alternatively, the new CEO may not want any of them around to stop him from achieving his own vision. Considering so many of key ex-ATI/AMDers executives are jumping ship, my guess it most of them either don't like the new CEO or are completely against the new direction he wants to go with AMD.

AMD's CEO has outlined that the focus for growth are emerging markets, markets with less competition, etc. It seems he is not interested in going head-to-head against Intel. He has also voiced concerns of transitioning to newer nodes more quickly or spending a lot of $ to compete in highly competitive sectors.

It's difficult to say how this will translate to the graphics division just yet. Right now GCN is pretty competitive but it was designed before RR was around by guys who loved graphics and envisioned holodecks by 2015-2016. If we are going to see a less competitive AMD in graphics, it won't be for a while. As NV has stated it takes about 3-4 years to design a new GPU architecture. Whatever GPU architecture AMD has ready to combat NV over the next 2 years has been designed before RR was at the helm. So AMD should be fine in the next 2-3 years. After that, it's hard to say. I still remain optimistic though since a part of their strategy is Fusion graphics. Therefore, they are going to need fast and efficient GPUs no matter what.

Their CPU side though I think is no longer on the radar for most of us. Once Haswell launches, Intel is probably going to open an even larger lead.

Nvidia stated that GPUs are no longer their primary focus.

NV's primary focus is still GPUs. Tegra is still a small fraction of their profits. I think you are misinterpreting the comment that NV made regarding new focus on high growth markets, which means tablets and smartphones. That's them addressing how to grow their business in other area, but it doesn't mean the bulk of their revenues and profits don't come from GPUs - they still do.

 

[mak360]

I am not sure I agree. NV is firing on all cylinders in GPUs. Kepler has more design wins in laptops than Fermi, their GTX680 just recaptured the performance crown in absolute performance, features, and key metrics AMD used to hold (Eyefinity no longer an advantage, performance/mm^2, performance/watt, performance/die). NV is going to launch full blown GK110 at some point. They are also working on launching Maxwell in 2014. It might seem like NV has slowed down but part of that stems from 28nm capacity issues and slower transition betwen new node shrinks. AMD is also not standing still. Sea Islands (GCN 2.0) is scheduled for Q1 2013.

It seems most ex-ATi/AMD employees are being rewarded with more lucrative job offers elsewhere. Also, it could be a lot of them are not aligned with the vision of the new CEO. Alternatively, the new CEO may not want any of them around to stop him from achieving his own vision. Considering so many of key ex-ATI/AMDers executives are jumping ship, my guess it most of them either don't like the new CEO or are completely against the new direction he wants to go with AMD.

AMD's CEO has outlined that the focus for growth are emerging markets, markets with less competition, etc. It seems he is not interested in going head-to-head against Intel. He has also voiced concerns of transitioning to newer nodes more quickly or spending a lot of $ to compete in highly competitive sectors.

It's difficult to say how this will translate to the graphics division just yet. Right now GCN is pretty competitive but it was designed before RR was around by guys who loved graphics and envisioned holodecks by 2015-2016. If we are going to see a less competitive AMD in graphics, it won't be for a while. As NV has stated it takes about 3-4 years to design a new GPU architecture. Whatever GPU architecture AMD has ready to combat NV over the next 2 years has been designed before RR was at the helm. So AMD should be fine in the next 2-3 years. After that, it's hard to say. I still remain optimistic though since a part of their strategy is Fusion graphics. Therefore, they are going to need fast and efficient GPUs no matter what.

Their CPU side though I think is no longer on the radar for most of us. Once Haswell launches, Intel is probably going to open an even larger lead.



NV's primary focus is still GPUs. Tegra is still a small fraction of their profits. I think you are misinterpreting the comment that NV made regarding new focus on high growth markets, which means tablets and smartphones. That's them addressing how to grow their business in other area, but it doesn't mean the bulk of their revenues and profits don't come from GPUs - they still do.

That’s a load of nonsense, show me the design wins, you can’t bcoz they are mostly Fermi wins (added onto Kepler). The mobile chips are mostly Fermi re-branded.

Again, you’re kidding right?

Hot air and rumours

You honestly think nVidia have less competition in ARM based chips LOL, goes to show you haven’t a clue what you’re talking about.
As someone here said “just Samsung & QUALCOMM, would eat nVidia alive”. NVidia has no chance vs these monsters; these guys don’t play games as nVidia does with AMD (marketing advantage).

 

[BlockheadBrown]

As always, Russian, thanks for your thoughts regarding this.

 

[Ventanni]

I think if anything, Intel will drive the graphics business in the next coming years; not because Intel will ever have a high end GPU, but because they'll finally have a successful, respectable low end one. The HD3000 is already pretty good for being an iGPU, Ivy Bridge is even better, and Haswell is supposed to be another huge improvement.

I guess what I'm saying is, unless AMD APUs actually have a superior graphical offering like they do now, there won't be *any* reason to ever use one. Intel CPUs are faster, more power efficient, and I'm pretty sure they can produce them at better yields than GF can. If Intel GPUs suddenly become a superior iGPU solution (far cry, I know, but Intel does have extremely deep pockets), then AMD is in trouble. So if anything, AMD is definitely going to be investing a lot of R&D into making sure they retain that competitive advantage.

 

[Olikan]

I think if anything, Intel will drive the graphics business in the next coming years; not because Intel will ever have a high end GPU, but because they'll finally have a successful, respectable low end one. The HD3000 is already pretty good for being an iGPU, Ivy Bridge is even better, and Haswell is supposed to be another huge improvement.

I guess what I'm saying is, unless AMD APUs actually have a superior graphical offering like they do now, there won't be *any* reason to ever use one. Intel CPUs are faster, more power efficient, and I'm pretty sure they can produce them at better yields than GF can. If Intel GPUs suddenly become a superior iGPU solution (far cry, I know, but Intel does have extremely deep pockets), then AMD is in trouble. So if anything, AMD is definitely going to be investing a lot of R&D into making sure they retain that competitive advantage.

actually i have more fear that NVIDIA will be out of the market sooner than AMD
With the igpu battle between Haswell and kaveri on the way, i have a fear that nvidia will lose every "power-efficient" gpu market,
theyr tegra 3 is a bulldozer compared to others A15 SOCs
they lost every console deal (i.e. they will have to pay for PhysX)

i am actually glad that kepler is a good arquitecture, Nvidia will need that

 

[blastingcap]

Nvidia stated that GPUs are no longer their primary focus.

What does it say about AMD if even a distracted Nvidia beats them? It's been three years now and AMD still has not fixed the tearing problem in Eyefinity. (Nv Surround does not have this problem.) It's like tessellation.. sure AMD got there first, but when Nv finally turned its attention to the problem, it was a clear victory.