Intel announced on Monday that it will be presenting a paper at Siggraph 2008 about its "many-core" Larrabee architecture, which will be the basis of future Intel graphics processors.
The paper itself, however, has already been published, and I was able to get a copy of it. (Unfortunately, as you'll see at that link, the paper is normally available only to members of the Association for Computing Machinery.)
Intel's Larrabee includes "many" cores, on-chip memory controllers, a wide ring bus for on-chip communications, and a small amount of graphics-specific logic.
(Credit: Intel)The paper is a pretty thorough summary of Intel's motives for developing Larrabee and the major features of the new architecture. Basically, Larrabee is about using many simple x86 cores--more than you'd see in the central processor (CPU) of the system--to implement a graphics processor (GPU). This concept has received a lot of attention since Intel first started talking about it last year.
... Read more
Finally, I can call myself an inventor.
I've been inventing things for almost 20 years now, but Montalvo Systems was the first company I worked for that took intellectual property seriously. (That was no coincidence; it was also the first company I worked for where I helped develop the intellectual-property strategy.)
During my years at Montalvo, I came up with quite a few ideas and participated in brainstorming sessions that yielded more ideas. Most of these sessions were limited to Montalvo's own people, but there was one person I brought in to help us as a consultant--Don Alpert, who was the principal architect of Intel's Pentium processor and, possibly less significantly, a member of the editorial board at Microprocessor Report.
Working with three of us from Montalvo--myself and chief architects Greg Favor and Peter Song--Don took the lead in preparing a set of related patent applications describing a new way to design microprocessors.
The first patent from this set was ... Read more
With Intel's focus on the new Atom-brand processors being described at the Intel Developer Forum this week, "Atom-powered" is the obvious description of the mobile Internet devices (MIDs) these chips will go into... and it seems like half the IDF stories on the Internet this week are using that phrase.
Intel's Atom processor (on the right) and its companion System Controller Hub code-named Poulsbo.
(Credit: Intel Corp.)Intel, however, seems to want even more hyperbole-- it expects people to believe that Atom will recharge the whole company. CEO Paul Otellini reportedly said "This is as important to Intel as the launch of the Pentium in the mid-1990s"-- but that's ridiculous.
The original Pentium processor and its descendants were responsible for nearly all of Intel's revenue. Atom will be merely a blip on Intel's financial reports.
The problem with Atom, especially these early models, is that the niche they occupy is a no-man's land between truly mobile devices like cellphones and MP3 players, and truly powerful devices such as laptop computers.
Atom consumes ten times as much power as cellphone processors and one-tenth the power of laptop processors. This power consumption makes for a device that has to be larger than a cellphone, and has to be smaller than a laptop because it can't provide comparable functionality. And there simply aren't enough applications that fit naturally into devices in that size range.
Look, I have as much experience with MIDs as anyone. I used an Apple Newton for seven years (as I've written about here several times). The Newton had roughly the same form factor and somewhat lower power consumption than today's MIDs. For a while I had a Metricom Ricochet wireless modem that gave me wireless Internet access.
But the simple fact is that the Newton wasn't useful enough to make me carry it around all the time. I loved mine because I had one critical application for which it was perfect. It was my electronic reporter's notebook, and no small-screen device could ever substitute for it. But most people don't need one of those.
And most people don't need a 5" to 7" display for basic Web browsing... at least, not enough to actually carry one around. And if you can carry something too large for a pocket, you can carry a small notebook PC that can handle a traditional notebook CPU.
Even after the Atom family evolves to the point that it can fit into cellphones-- which is the only way it's going to achieve significant sales volumes-- profits from these chips will never be very high. Intel's never going to achieve a monopoly in cellphone processors, and the competition from ARM-based cellphone chips will keep the value of a CPU core under a few bucks.
What Intel doesn't want people to think about very much right now is that in a cellphone, the CPU core is about the least-valuable part of the system. Even in a MID with an Atom processor, the CPU is just a tiny part of the whole package. Look at that picture up there-- the Atom processor is small compared with its companion system controller. In a cellphone, there's even more circuitry required for the radios.
And that's why Intel's never going to be an Atom-powered company, and I'm sure Otellini knows that in spite of everything he's been saying. But when your stock price has been trending downward for seven years in spite of the fact that you're running the world's largest semiconductor company with a stranglehold on the world's largest semiconductor market, I suppose you have to try to drum up as much excitement as possible for every new product that comes along.
- prev
- 1
- next
