Beware the Marketing Department’s Re-branding Efforts
Intel and AMD have just released some hot new processors - the 2nd Generation Core processors and Fusion, respectively - and I think they're both significant steps forward. Unfortunately, they're also both mired in some marketing jargon and branding schemes that seem intentionally designed to make users think they're getting something they're not. Or at least, to confuse them into not knowing exactly what they're getting. Sometimes I think the folks in the marketing department come up with naming and branding schemes that aren't necessary just to justify their paychecks.
Let's start with Intel. The processors formerly code-named Sandy Bridge have been dubbed the 2nd Generation Core Processor Family and still carry the Core i3, Core i5, and Core i7 monikers. Really, they're not the 2nd generation of the Core i3, i5, and i7 processors we've known for a couple years. The first were the Lynnfield and Bloomfield processors made using 45nm process technology. The 2nd generation was the move to a 32nm manufacturing process with the Gulftown, Clarskfield, and Arrandale processors. The microarchitecture used in these former was called Nehalem, and was tweaked in the 32nm generation to Westmere. It wasn't just a shrinking of the manufacturing process technology, either: Westmere added instructions to accelerate AES encryption, improved virtualization performance, and in some models integrated graphics to the processor package (though not the actual same piece of silicon as the CPU).
What Intel calls 2nd Generation is really a wholly new microarchitecture, still on a 32nm process. It's a re-design of the CPU from the nitty gritty insides of the CPU cores themselves. It's more efficient, less power hungry, and is loaded with new features, from AVX instructions to a ring-bus communication system to the new, more powerful graphics processor built right onto the CPU die itself. If Intel wanted to be clear, it would have chosen a new name for what is really a very new processor, or at least called them Core i4, Core i6, and Core i8 or something similar. Instead, there's re-jiggering of the model numbers. We wouldn't want the average customer to stop buying the old generation processors just because they could clearly identify them, right?
Intel's also trying to re-brand its integrated graphics as "processor graphics", to make the distinction that the GPU is actually built into the processor die itself. While that is certainly new, let's not kid ourselves: it's still integrated graphics. It's literally more integrated than ever, and still shares memory and power with the CPU.
AMD's "Discrete Class" Graphics and APUs
Intel's insistence on re-branding its new integrated graphics isn't as bad as what AMD is doing with its new Fusion E- and C-class processors. The company and its partners refer to the Radeon HD 6310 and 6250 graphics in these processors as "discrete-level" or "discrete-class". While the graphics part of these processors share a common design architecture with the Radeon HD 6000 line and supports DirectX 11, it is by no means "discrete-class". The integrated graphics in these Fusion processors is really quite fast for such inexpensive, low-power CPUs, but even the least expensive modern graphics cards are quite a lot faster. The processors still share power and memory with the main CPU - that's integrated graphics any way you slice it.
AMD has gone so far as to claim its new Fusion CPUs aren't really CPUs, they're "APUs". Besides making me think of the lovable convenience store clerk from The Simpsons, this is just a made-up new class of product that only AMD seems to think exists. The Fusion parts are APUs (Accellerated Processing Units) because they put the CPU and GPU and video processing all on the same die. That GPU and video processor can help accelerate tasks, like video transcoding, that bog down the CPU. Of course by that definition, Intel's new Sandy Bridge processors are APUs as well. So are most modern smartphone system-on-chip designs. The thing is, I actually think this APU designation might make sense, if we lived in a world where Intel, ARM, Nvidia, and others were all behind it. But a category that exists only for a single vendor, when other vendors have products that qualify, is no category at all.
Nvidia's New Generation of Old-Generation GPUs
Lest I leave Nvidia out of the "let's send Jason Cross an angry email from the marketing department" list, I have a bone to pick with the products naming for its graphics cards. The GeForce GTX 580 uses a chip that is essentially the same one found in the GeForce GTX 480. It's full of transistor-level fixes to improve performance, but the design is exactly the same. Typically, one would think that the jump from the 4xx series to the 5xx series would indicate a new design, not simply a faster version of the old design. Nvidia has done this a lot. The GeForce GT 230 and GTS 240 were not based on the GT2xx series of GPUs that the rest of the GeForce 2xx cards were. Rather, they were based on the G92 chip that the GeForce 9800 cards were. Many of the GeForce 300M mobile graphics products were based on the same exact chips as the GeForce 200M products, just running at different clock speeds. The lower end of the GeForce 500M mobile series are based off the same exact GPUs as the GeForce 400M line.
Nvidia has a storied history of taking the last generation's GPU, turning up the clock speeds, and slapping a newer generation name on it. This doesn't make any of these products bad, mind you. In many cases, these are quite excellent graphics cards. It's just a little disingenuous to suggest a newer generation of product based on the name, when the technology you get is from the previous generation.
Isn't the tech world confusing enough? Do manufacturers really have to muddy the waters with confusing terms and products names that make it less clear what you're really getting? Intel, AMD, and Nvidia aren't alone in this. Consider the Motorola ATRIX phone and Xoom tablet. Motorola says these are 2 GHz devices, blatantly misleading consumers. They each contain a dual-core 1 GHz processor, and Motorola is just adding them together. This is a load of bull; the quad-core 3 GHz processor in my desktop PC is by no means a 12 GHz CPU. Unfortunately, I don't expect this sort of creative manipulation to stop anytime soon (just look at the whole "we have renamed our 3G cell network to 4G" debacle going on now). As a technology consumer, you'll just have to stay on your toes, and ask independent experts like your friendly PCWorld editorial staff to help you make sense of it all.