I don't write much about hardware, but every once in a while a hardware advance comes along that is so important that it deserves some attention. There are pragmatic reasons for this. What Bill Gates gratingly called "the magic of software" is the driving force behind our industry—the secret sauce that makes all these devices we interact with, PCs or otherwise, more interesting.
But this year, we begin a pivotal hardware transition that's as big a milestone, I think, as previous transitions. There's just one problem: Unlike the past, where we could simply point to a single thing—the move from 16-bit processing platforms such as the 286 to a 32-bit world of flat memory models—this year, it's a bit different. That is, it's not a single transition. It's a series of transitions.
The first is simple enough, however. Intel recently launched its "Sandy Bridge" microprocessors, which are the second generation of its Core-branded chips. But despite the identical naming—you'll still be picking between Core i3, i5,and i7 variants—these new designs are much improved over their predecessors.
As before, there will be dual- and quad-core variants of the chips, but this time around there will be far more of the quad-core versions, especially for desktop PCs. Step up to the i5, and you'll find some 8-core versions, and with the high-end i7, even some 12-core chips. (These latter chips have not yet shipped.)
The big deal about Sandy Bridge, aside from raw performance, is that it marks the first time Intel has integrated so much technology onto a single chip, marking the microprocessor giant's biggest redesign since the Pentium 4. So this time around, there's a high-performance integrated graphics processor, shared L3 cache, a memory controller, and PCI Express links. What this means to you is that Sandy Bridge-based PCs will achieve better power management than ever before, and, for portable machines, far better battery life. And this comes in addition to that performance boost.
I happened, coincidentally, to skip the first-gen i3/5/7 chips, my fastest computer being an older quad-core Core 2 Quad-based affair. But now I'm glad I did: PCs based on Sandy Bridge are going to make last year's models look sad by comparison.
ARM and SoC
Microsoft caused a lot of excitement at the Consumer Electronics Show (CES) earlier this month when it announced that the next Windows version, called Windows 8 currently, would run on system-on-a-chip (SoC) designs in addition to standard PC-based x86/64 systems. Among the supported SoC designs, too, would be those based on the ARM platform.
This announcement confirmed previous rumors about such a port, but Microsoft also threw a few mysterious wrenches into the mix as well. First, even though my sources at the software giant continue to insist that Windows 8 will ship by "late 2012," Microsoft's Windows lead, Steven Sinofksy, said at the announcement that the system was "two to three years" away, so it's unclear if he meant Windows 8 on SoC generally, Windows 8 on ARM, or just Windows 8.
Second, Microsoft CEO Steve Ballmer asserted that he saw Windows (8) as the future for all of the company's mobile computing efforts, be them traditional PCs (like notebook computers and tablet PCs) or other, upcoming devices that are smaller and even more portable. This of course raised questions about the future of Windows Phone, since some of these coming devices will surely be smartphones.
Regardless of how the details of this play out, what's happening here is important: Microsoft is embarking on a new generation of its "Windows Everywhere" strategy , and this time it appears to be playing hardball: The company actually showed a native port of its Office suite running on ARM, for example. Point being that Windows on ARM is Windows, and not just something that looks like Windows.
(I happen to disagree with this strategy. See Windows Everywhere? Wake Up, Microsoft! It's 2011! for details.
iPad and Other Tablets Point the Way
What Sandy Bridge and SoC have in common is intense hardware integration. But taking things a bit higher level, we see another important trend that will have huge implications this year. Apple will deliver a second-generation iPad, one that will no doubt remove all the issues I and other critics have complained about. And Apple's competitors, including those allied around Android, plus HP with webOS and RIM with the PlayBook, are set to test the market with a huge new range of iPad-like devices.
Apple's new iPad will be based on some future version of the company's A4 chip. Other tablets will use similar designs, including the NVIDIA Tegra 2. What these things all have in common is superior battery life, multi-core performance, and the same type of SoC design that Microsoft is only starting to embrace for future products.
Where 2010 was the year of iPad's birth, however, 2011 will be the year that the iPad—and perhaps some of its competitors—truly flourish. And what's interesting about that is that these devices are not PCs, not even Macs, but rather a new type of device that has much in common with smartphones. And that's not a bad thing: The systems that drive these devices are simpler than PCs.
So 2011 will be seen as the year that everything changed. PCs will become more powerful and yet smaller and more energy efficient. Users will migrate in ever bigger numbers to simpler, non-PC computing devices, and that trend will only accelerate going forward. And all of this is happening due to the same types of hardware advances happening on different platforms.