What Took So Long?

AMD announced the acquisition of ATI in 2006. By 2007 AMD had a plan for CPU/GPU integration and it looked like this. The red blocks in the diagram below were GPUs, the green blocks were CPUs. Stage 1 was supposed to be dumb integration of the two (putting a CPU and GPU on the same die). The original plan called for AMD to release the first Fusion APU to come out sometime in 2008—2009. Of course that didn't happen.

Brazos, AMD's very first Fusion platform, came out in Q4 of last year. At best AMD was two years behind schedule, at worst three. So what happened?

AMD and ATI both knew that designing CPUs and GPUs were incredibly different. CPUs, at least for AMD back then, were built on a five year architecture cadence. Designers used tons of custom logic and hand layout in order to optimize for clock speed. In a general purpose microprocessor instruction latency is everything, so optimizing to lower latency wherever possible was top priority.

GPUs on the other hand come from a very different world. Drastically new architectures ship every two years, with major introductions made yearly. Very little custom logic is employed in GPU design by comparison; the architectures are highly synthesizable. Clock speed is important but it's not the end all be all. GPUs get their performance from being massively parallel, and you can always hide latency with a wide enough machine (and a parallel workload to take advantage of it).

The manufacturing strategy is also very different. Remember that at the time of the ATI acquisition, only ATI was a fabless semiconductor—AMD still owned its own fabs. ATI was used to building chips at TSMC, while AMD was fabbing everything in Dresden at what would eventually become GlobalFoundries. While the folks at GlobalFoundries have done their best to make their libraries portable for existing TSMC customers, it's not as simple as showing up with a chip design and having it work on the first go.

As much sense as AMD made when it talked about the acquisition, the two companies that came together in 2006 couldn't have been more different. The past five years have really been spent trying to make the two work together both as organizations as well as architectures.

The result really holds a lot of potential and hope for the new, unified AMD. The CPU folks learn from the GPU folks and vice versa. Let's start with APU refresh cycles. AMD CPU architectures were updated once every four or five years (K7 1999, K8 2003, K10 2007) while ATI GPUs received substantial updates yearly. The GPU folks won this battle as all AMD APUs are now built on a yearly cadence.

Chip design is also now more GPU inspired. With a yearly design cadence there's a greater focus on building easily synthesizable chips. Time to design and manufacture goes down, but so do maximum clock speeds. Given how important clock speed can be to the x86 side of the business, AMD is going to be taking more of a hybrid approach where some elements of APU designs are built the old GPU way while others use custom logic and more CPU-like layout flows.

The past few years have been very difficult for AMD but we're at the beginning of what may be a brand new company. Without the burden of expensive fabs and with the combined knowledge of two great chip companies, the new AMD has a chance but it also has a very long road ahead. Brazos was the first hint of success along that road and today we have the second. Her name is Llano.

The Llano A-Series APU


View All Comments

  • ET - Tuesday, June 14, 2011 - link

    So, what do you tell them? The only benchmarks of The Sims 3 I could find are pretty old and didn't offer much detail, but I think based on them that high quality would require more than the lowest IGP. (Then again, normal or low quality should probably run fine on anything.) Reply
  • msroadkill612 - Tuesday, August 02, 2011 - link

    ta 4 the post - but dont salesmen have a duty to sell them a bit of insurance against the next game fad Reply
  • ash9 - Tuesday, June 14, 2011 - link

    Bring on the OpenCL apps, excuse me applications- Excel rewritten to take advantage of heterogeneous computing would silence everyone about Star CPU cores. The ball is in the hands of the people (to buy them), then the software developers (to program for them) - that's why Star cores? APU apps needed Reply
  • msroadkill612 - Tuesday, August 02, 2011 - link

    I think it will happen. Big mobs will identify niches that can profit from OGPL & profit from using it - open languages always win in the end.

    I see fusion server apuS in the future.
  • FISHRULE - Tuesday, June 14, 2011 - link

    What a terrible CPU, who would want something that performs like a Phenom in a new computer circa 2011. The future might be fusion AMD, but you sure as heck aren't part of the future anymore. Reply
  • jabber - Tuesday, June 14, 2011 - link

    90% of laptop owners wouldnt know if they were using a Brazos or an i5 in their laptops.

    The only real differential is in transcoding, ripping etc. and very few folks in the real world actually do that. Especially on laptops.

    Price is far more important than outright performance. Has been for some time now.
  • RussianSensation - Tuesday, June 14, 2011 - link

    False. My gf doesn't know anything about computers. But she somehow knew that Intel makes the best mobile CPUs even before asking me what to get. Clearly Intel marketing > AMD's. All of my current friends who just bought a computer all went with i3/i5/i7 processors only because they "heard" Sandy Bridge is the fastest CPU around.

    Llano is nothing more than a Phenom with a faster GPU. Phenom already didn't sell well against C2D/C2Q/Core i7 (1st) gen and isn't getting any better against SB. The only way AMD is gaining market share is if they ship cheap laptops with Llano to users for whom the price of a laptop is the most important factor.
  • ET - Tuesday, June 14, 2011 - link

    Well, you have pretty computer literate friends. Most people I know would have no idea what Sandy Bridge is, probably not even what i3 and i5 are, and would only buy i3 or i5 because: a) AMD had very little market presence until now; b) I'd recommend them. That said, most of the people I know have no idea that AMD exists and is making CPU's (I mention it occasionally, and they're always surprised, so I guess it doesn't register), so it does look like AMD's marketing is pretty crappy.

    That said, I think that your analysis of why Llano will fail isn't right. For most people the CPU power really doesn't matter that much. They'd have no idea if i3 or Llano is better, and most likely won't be able to tell the difference in practice (unless they run a game for which the HD 3000 is unsuited). Sure Llano is for the low end market, but that's where most sales are, and it's certainly much improved in terms of power usage, which is an important enough measure to help it gain market share.
  • RussianSensation - Tuesday, June 14, 2011 - link

    You are probably right that most people won't tell the difference between a Llano or an i3/i5/i7. But to them the perception of knowing that they have a slower CPU is what matters. One of my friends was building a PC for browsing the net only. I told him he'd be perfectly fine with a $100 CPU but he insisted that the system must have Sandy Bridge because it's the latest modern CPU. I gave up trying to convince him that his internet browsing experience will be more limited by his ISP latency and speed rather than CPU performance.

    So as long as AMD convinces the average consumer that Llano is at at least as good, they will do well. The problem is AMD's marketing department is worse than a 1st year undergrad student studying marketing in business school. They think if they pay millions of dollars to put AMD on F1 cars, people will notice?

    You made a very important point - a lot of people don't even know what AMD is or that there is another competitor to Intel. Imagine if GM, Ford or Chrysler made cars that were more reliable than Honda or Toyota. It would still be a while until the average consumer would abandon the Japanese brands since the perception of reliability would lag behind reality. AMD has this similar problem with their CPU brands, which only marketing can fix.
  • jrs77 - Tuesday, June 14, 2011 - link

    Really... that's rather unimpressive.

    The GPU of Llano beats intels HD-graphics in games, what a surprise. But how many people do I know personally, that play games on their laptops anyways? Laptops are primarily used as mobile office-computers and they do still rely more on CPU-performance.

    From all the experience I've made the HD3000 graphics of mobile SNB CPUs are perfectly fine for all tasks I throw at them (excluding games). So the question is, why would I buy a Llano-based laptop instead of a SandyBridge one?
    Llano doesn't offer better battery-life so the only reason might be the price, but with i5-2xxx laptops starting at $600 I really don't see alot of competition there for intel, if we're talking anything else then gaming.

Log in

Don't have an account? Sign up now