How would you like a single-chip microprocessor with more than four times the performance (on some applications) of Intel's best Core i7?
Then consider that up to 32 of these chips can be directly connected to form a single server, achieving four times the built-in scalability of Intel's next-generation Nehalem-EX processor.
That's IBM's widely anticipated Power7, which it described at last week's Hot Chips conference. But if you're interested, you'd better be prepared to spend a lot more than four times as much per chip. IBM isn't talking about pricing, but large Power servers can cost more than $10,000 per processor.
IBM's forthcoming Power7 server processor has eight cores, manages 32 threads, and includes 32MB of on-chip embedded DRAM cache. Power7 also has the highest levels of off-chip bandwidth ever achieved by a microprocessor.
(Credit: IBM)What makes the Power7 so powerful? Each chip has eight cores, and each core supports four-way multithreading. There's 32MB of level-3 cache on the chip, made using embedded DRAM (eDRAM) cells. Most CPUs use SRAM for cache because it's generally easier to combine with high-performance logic, but DRAMs--with only one transistor per bit--offer compelling density advantages. IBM spent years developing a new kind of eDRAM that would work with SOI (silicon on insulator) manufacturing processes, and the Power7 is the most advanced product to use the new technology.
Interestingly, the Power7 cores run much more slowly than those in the Power6 processor, which I wrote about here in 2007 ("Live from Hot Chips 19: Session 1, IBM's Power6"). The Power6 was designed to run very fast using a long CPU pipeline in order to deliver the highest possible performance on each thread of execution.
Maybe that strategy didn't work out as well as IBM hoped, because the Power7 returns to a more traditional microarchitecture with a shorter pipeline and much lower clock rates--though IBM didn't say exactly what those rates would be.
IBM did, however, promise that the Power7 would be roughly four times as fast as the Power6, chip for chip. Since it has four times as many cores, each of the new slower-clocked cores must still deliver about as much performance as those in the previous generation.
Chip-level performance must always be matched by off-chip connections lest the incoming data or outgoing results be bottlenecked by a too-slow channel. Accordingly, the Power7 is equipped with eight I/O channels for DRAM, each of which connects to an off-chip buffering device that splits the channel into two 64-bit DRAM interfaces. All together, IBM says the Power7 has 180 GBps of DRAM interconnect that can sustain over 100 GBps of effective memory bandwidth.
There's another 50 GBps of peak I/O bandwidth and a staggering 360 GBps of peak bandwidth used to let each Power7 chip communicate with others. The DRAM connected to each chip is thus shared across larger systems.
Combining these figures, IBM says a single Power7 has 590 GBps of total off-chip bandwidth. This isn't the real number, since many of those bytes are used for error-correcting codes and other overhead, but it's still pretty impressive.
So is Power7's die size: 567 square millimeters for 1.2 billion transistors. That's nearly a square inch! IBM says that if the 32MB L3 cache had been manufactured using SRAM, the transistor count would have been 2.7 billion instead.
Still, Power7 wasn't the only high-end chip talked about at Hot Chips.
Rainbow Falls, a record for core count
Sun Microsystems was there to describe its forthcoming Rainbow Falls chip, which I assume will be marketed as the UltraSparc T3. The chip has 16 cores, each of which is reportedly able to manage 8 threads.
Sun's primary Rainbow Falls presentation focused on details of Rainbow Falls' internal and external interconnects; a second talk described the cryptographic coprocessors present in each of the chip's cores. These coprocessors--one for modular arithmetic (commonly used in public-key cryptography) and a cipher/hash unit to accelerate bulk ciphers like AES and secure hash algorithms--provide many times the performance of pure software implementations.
Fujitsu was also at Hot Chips to describe its eight-core, 2GHz Sparc64 VIIIfx processor, the latest in a long series of impressive designs from the company. Fujitsu quoted a peak performance figure of 128 GFLOPS (billions of floating-point operations per second) with a typical power consumption of just 58 watts. It did not, however, provide sustained performance or worst-case power consumption figures.
AMD, Intel vie for high-volume servers
Few of us will have direct exposure to the IBM, Sun, and Fujitsu chips. A pair of presentations from Advanced Micro Devices and Intel described products that will be much more widely available.
AMD launched its six-core Opteron processor code-named "Istanbul" earlier this year (see Brooke Crothers' coverage from June). Next year the company will begin shipping a new Opteron model currently code-named Magny-Cours (after a racetrack in France). Magny-Cours will consist of two Istanbul chips in a single package, with twice as many DRAM interfaces to support the new processor's increased performance.
AMD also teased the audience with another mention of a new processor core design that has been under development there for several years: "Bulldozer," which is now targeted at 32nm process technology. This new core will incorporate new x86 instruction-set extensions which will probably not be adopted by Intel (a strategy that reminds me of AMD's old 3DNow extensions).
But saving the best for last--best, that is, from the perspective of anticipated sales--Intel's talk on Nehalem-EX showed just how far Intel has been able to push the technology envelope for high-volume servers.
Nehalem-EX is an eight-core version of the existing quad-core Nehalem design. The new chip also has 24MB of L3 cache done in old-school SRAM. By my calculations, about 60 percent of the chip's 2.3 billion transistors are in this cache alone.
Nehalem provides four links to external DRAM buffer chips supporting two DDR3 DRAM interfaces each (much like the Power7 solution) and four QuickPath Interconnect links that provide direct "glueless" connections for up to eight-processor systems (64 cores, 128 threads). Intel is also working on an external Node Controller chip for systems with up to 2,048 Nehalem-EX processors.
The aggregate bandwidth numbers for Nehalem aren't as mind-boggling as those for Power7, but they're still far beyond anything available for PC-architecture servers today. Based on the presentation, I estimate Nehalem could boast over 85 GBps of peak memory bandwidth and 100 GBps of chip-to-chip bandwidth, some of which must be allocated to I/O.
I expect the raw number-crunching performance of the Nehalem-EX cores to be roughly on the same level as Power7's cores. The lower ratio of bandwidth to processing power for Nehalem-EX reflects a different design target, not a design shortfall--and most importantly, a much lower selling price. There will presumably be versions of Nehalem-EX priced similarly to existing Xeon MP products, which currently top out at $2,301 each in small volumes, but that's a very reasonable price to pay for the market's most advanced x86 server processor.
I suppose if I were just in search of controversy, I'd write a post to proclaim the death of the MID (mobile Internet device) category. My obituary for the Netbook earlier this week generated a ton of traffic; I suppose I could do that again. Certainly, the concept of a MID--a device midway in size and capability between smartphones and the smallest notebooks--is under tremendous pressure from both sides.
Customers have learned that with a well-engineered browser, the small displays on phones such as Apple's iPhone and T-Mobile's G1 "Google phone" are sufficient for most Internet applications (Web browsing, e-mail, chat, etc.). And as I described yesterday, small notebooks are quickly lifting themselves out of the "Netbook" ghetto, gaining performance and cutting power consumption to become reasonable alternatives for those times when a smartphone just isn't enough.
The tokidoki edition Fujitsu LifeBook U820 mini notebook.
(Credit: Fujitsu Computer Systems)But I think there's still a legitimate niche for MIDs and other miniature mobile PCs. As I've mentioned here before, I used to carry around a 1.5-pound computing gizmo along with a conventional laptop. It was an Apple Newton MessagePad 2100--officially a PDA, not a MID--but it was as close to a MID as the technology of the time allowed. It came with a Web browser, and for a while I had mine equipped with a Metricom Ricochet wireless modem, so I could access the Web and e-mail on the go.
It often seems to me that I would like to go back to that kind of device, rather than trying to make my iPhone and my laptop do the same jobs. In fact, I think my note-taking capability has actually declined with each new handheld platform I've adopted--the Newton was better than the Palm Treo, and the Treo was better than the iPhone. Today, when I attend conferences or want to scribble down some idea that can't be represented in a paragraph or two, I grab a Moleskine notebook (the pocket Sketchbook version).
My own experience is merely anecdotal evidence, however, and I know better than to rely on that. So what are the real markets for the MID?
Coincidentally, I think it works out to three E's: education, entertainment, and executive applications. All three areas lead to situations where a person might want access to more computing and communications resources than a smartphone can provide but won't necessarily want to carry around a notebook--or try to use one while standing--to get that.
The educational market for these small machines has yet to develop because current MIDs don't yet offer the right combination of small size, all-day battery life, and low price, but I believe they'll get there within the next year or so. People often talk about e-book readers as being the right answer for educational computing, but e-books are more about static content, and education is ideally an interactive process.
The entertainment focus was clearest with UMPCs (another dead category, though I'm hardly the first to point that out). UMPCs were marketed as "lifestyle" gizmos, as if many people were ever going to make a relatively bulky 7-inch display tablet PC with two-hour battery life part of their lifestyle. But in a smaller form factor--say a 5-inch display, a total weight under a pound, and battery life of at least five or six hours--a MID can fit this bill. As long as it's small enough (and rugged enough) to carry around in a purse or jacket pocket, and cheap enough to be written off to the entertainment budget like a Netflix subscription or a new TV, a MID could indeed become a lifestyle product.
The Viliv S5 Entertainment MID provides full PC compatibility in a PDA-size package.
(Credit: Yukyung Technologies)I saw a gizmo at CES that fit this definition pretty well, the Viliv S5 from Korean consumer-electronics maker Yukyung Technologies. Yukyung is one of many companies making portable video players, but its new offerings are quite distinctive.
The S5 is like a right-sized UMPC, with a 4.8-inch touch-screen display (800x480 or 1024x600 pixels, depending on model). It can play HD video, and it comes with Windows XP on a real hard disk, so there's no problem installing other software.
The S5's Intel Atom processor provides very good battery life: the company specifies six hours of movie playback. The device is about 6 x 3.3 x 1 inches in size--a lot smaller than my old Newton--and weighs less than 14 ounces.
There are also two 7-inch screen Viliv machines, the X70 slate-style tablet and the S7 convertible tablet. Both, amazingly, are still smaller than my old Newton.
Executives have always been the focus of some high-end handheld PC developers such as OQO, Sony, and Fujitsu.
Fujitsu didn't have any major updates to announce at CES for its LifeBook U820 series, though it was showing a model with case art from tokidoki, an Italian (but Japanese-inspired) lifestyle brand, and I got a chance to talk with a couple of PR people from Fujitsu about the U820 and other Fujitsu products.
The U820 is basically a complete convertible tablet PC squeezed into a 1.3-pound package: a 5.6-inch touch-screen LCD with 1,280x800-pixel resolution, a 1.6GHz Atom processor, 1GB of RAM, a 60GB or 120GB hard disk, Windows Vista Home Premium, and so on. It offers pretty much every kind of communication technology a person could ask for: Bluetooth, a/b/g/n Wi-Fi, optional AT&T wireless broadband, and even a GPS receiver.
From my perspective, the U820 is actually smaller than it needs to be, which is most apparent in the micro-sized keyboard, but it's an impressive technical accomplishment nonetheless.
For many people, the new Sony Vaio P-series (a CNET Best of CES award winner this year) may prove to be more practical, with its 87 percent-pitch keyboard and 8-inch widescreen LCD. But the Sony is beyond all but the largest pockets. Sony has made smaller machines in the past, such as the Vaio UX series, but these have been discontinued.
The OQO model 2+ brings better performance at a lower price than earlier OQO models.
(Credit: OQO, Inc.)OQO also made a big splash at the show with its new model 2+, an unprepossessing name for a product even more technically impressive than Fujitsu's. The new OQO machine has almost all the features of the U820, but in a considerably smaller, lighter package. There are some differences; the model 2+ has a lower screen resolution (800x480) but is available with a faster CPU and more RAM. Also, the OQO is available with an OLED (organic light-emitting diode) display that really looks fantastic, with high contrast and deep saturated colors.
The model 2+ is in the same enclosure as the older OQO model 2, hence the trivial name tweak, but there's another big difference from that older product: the 2+ has a starting price of just $999, $500 less than the starting price of the 2. And the base model of the 2+ is a much better system than the high-end model 2 configuration was.
Just as there were some ARM-based Netbooks at CES, there were also some ARM-based MIDs on display. With no clear advantages over smartphones except for display size, I don't think these products will attract customers. But that problem is CPU-specific; it doesn't apply to the more powerful x86-based products.
So okay, there's some good MID hardware out there. Unfortunately, that isn't enough. What MIDs need are lower prices, more rugged designs, and some MID-optimized software. The fact that Windows runs on these small displays doesn't mean that style of user interface is right for them. I know people at Microsoft who are working on this aspect of the problem; I hope they get the chance to bring their solutions to market, ideally in the Windows 7 time frame.
All in all, there's a lot of interesting activity in these smaller form factors. I think these tiny machines face a long uphill struggle to gain market share, but at least they have a unique and clearly defined product concept: a PC in a pocket.
- prev
- 1
- next
