Back in 2009 I first wrote about ARM in a rambling I titled ARM's Ascension in which I talked about the rising aspirations and potential of ARM processors in the general computing field. A lot of what I said still rings true, and a lot of what seemed apparent in the future of the market back then is now known to not be true anymore. Smartbooks were prototyped many times, but never made it into shipping products. Instead what happened was the iPad.
iPad ended up doing exactly what many other Apple products have done in the past. When Apple entered the portable media player market, it flourished. When they entered the smartphone market, it flourished. Now that they've entered the tablet market, or maybe better said, initialized the tablet market, that market is set to flourish also. At the forefront of this new emerging form factor is ARM. No matter what SoC your product is using, be it Apple A4, Qualcomm Snapdragon, Samsung Hummingbird, or NVIDIA Tegra 2, ARM lies in the center of it. From the get-go it seems ARM has an iron grip on the market, leaving competitors, namely Intel, with a cliff face of an uphill climb if they want in on it.
The problem with Intel is that they're acting complacent, even arrogant, towards the emerging ultra portable trends. Atom was a good initial first step in that direction, with the ultimate goal being smartphones, but Intel failed to execute on their own plan to put that into motion. In essence, the crux of the success of Atom rested on Intel's superior manufacturing advances. Intel was a generation ahead of their nearest competitors on process size, and Atom would have been a great testing ground for their 32nm process launched a year ago. Instead Atom is still stuck on the two year old (plus) 45nm process, allowing competitors like ARM and AMD free reign on low-power alternatives built on more efficient technology. This complacency is what drove Microsoft to announce the biggest shift in direction for Windows since its inception.
For over 20 years, Windows was built off of the basis of x86 architectures. This was initially necessitated by the original foundation on which it was built off of, MS-DOS. In the span of Windows' existence, several offshoots of the core desktop and professional products have been spun for smaller, chiefly embedded devices, but those were almost always underdog products in a world that was dominated by proprietary solutions. The latest offshoot, Windows Phone 7, has seen mild success, but rather than enhance that OS and port it over to larger devices, Microsoft wanted to spread the reach of their bigger, mainstay OS into smaller devices. Teaming up with Intel, Microsoft was able to keep Linux at bay in the netbook market, and no doubt the plan was to continue riding Atom's coattails into smaller form factors as Intel continued to refine and slim its power characteristics. That did not happen. Instead Intel got, for lack of better reasoning, lazy, and Microsoft seemingly got fed up waiting. So they announced Windows 8's support for ARM processors.
This is huge. First of all it breaks the public perception of the alliance Intel and Microsoft have always had. Secondly it takes away one of the key advantages of Intel going into smaller form factors. Without being able to keep Windows all to themeselves, Intel can't tout the advantage of industry standard x86 support, primarily in regards to one of the biggest pieces of x86 software there is. Windows tablets with Intel and x86 will still exist, but their claim to fame will be PC software compatibility at the cost of ARM-type power efficiency and battery life. Arguably x86 compatibility doesn't mean much in ultra-portable devices, and even Intel will be quick to point out their support for Linux type lightweight OSes, such as Moblin. Additionally, though the possibility does remain slim, there is the remote chance that Google could announce x86 support for Android, effectively nullifying whatever blow to the company Microsoft's announcement may have had. But of course, Google seems perfectly content where they're at, and looks to be shaping up to be the OS provider of choice for smartphones and tablets not branded with pomaceous fruit.
In fact it's been confirmed that Google's platform of choice is one ARM solution in particular, NVIDIA's Tegra 2. NVIDIA's been getting a lot of good press lately, raising hope for their future and (not inconsequently) causing their stock price to shoot up. Biggest among them is Project Denver, an actual enthusiast, high-performance ARM processor. That alone would have been pretty neat, but coupled with Microsoft's timely revealed support for it, gives the prospect an incredibly tantalizing appeal. There's almost enough just in that one nugget to dedicate a whole rambling to, but there's tons of nuggets to digest this month, so I'll just say that while the notion is extremely interesting, NVIDIA will have a struggle getting not only the performance in place, but the software support (remember how well CUDA went?), and even if either of those factors are squared away, they still have to somehow differentiate their product enough to sell not just gamers this time, but average consumers on its potential. And since I've mentioned gaming, how will their graphics cards work with this new CPU? Will they come with PCIe slots, or will it exclusively work off of integrated graphics on the processor itself? And furthermore, without legacy PC game support, or presumably game support from future big-budget franchises, how will they justify their graphics performance advantage that has been the bread and butter of the company since its founding? Will they simply port Android games made for Tegra 2? Tons of questions loom over this announcement, more than we have answers for, and only time will reveal what those are.
The software question extends not just to games but in general. ARM has always been eyeing the specter of bigger PCs, only daring to dream the dream in recent years as their processors have been catching up to entry level laptop computers (read: netbooks). ARM finally has a backer willing to realize that dream in NVIDIA, but will the much-touted advantages of ARM's efficiency carry over into larger systems where transistor counts and razer thin power envelopes aren't such a concern? While the idea is fascinating, Intel's ironclad hold on the desktops and full-sized laptops seems impenetrable, not only in regards to the install base but also in the sheer vastness of the software support, with few consumers likely willing to move away from it, probably the most staunch of whom are the enthusiasts themselves. Average Joes may not care what kind of system they play their blu-ray movies on, but tech savvy users might have a whole catalog of applications that they've grown accustomed to over the years. It's a fun dream to ponder over, but for now ARM probably doesn't hold out much hope beyond the ultra-mobile stronghold they've already established.
An ultra-mobile market that's still very much evolving, I might add. When netbooks first came out, there was a lot of experimentation before they eventually settled on the 10" form factor (with some deviations into 12" sizes). This was chosen for the readability of the screen, as well as the size of the keyboard. Whatever device you're using, a key requirement is the ability to comfortably type. Tablets are riding on the expertise gathered from the netbook market, but as it is, they're fairly awkward devices to use. Just browsing is generally fine, but if you have to type out something relatively lengthy, you'll have a hard time finding a comfortable seating position. It either has to lay flat on your lap, or flat on a table, either way placing the screen at a somewhat extreme angle, either necessitating an IPS panel type, or straining your neck. Holding the device while typing isn't a very pleasing option, because you're either stuck typing one-handed, or trying to use your thumbs, for which most tablets are too wide. 7" tablets are a small majority of the upcoming offerings, and are only really good for thumb typing. But then 7" tablets are too bulky to carry around everywhere you go, neither fitting in your pocket or sitting very comfortably at your hip with a belt holster. So you're stuck using it at home, which seems kinda pointless with such a small screen given all the alternatives you'll likely have available. Smartphones are great for providing the full web experience everywhere you go, but they're not quite the decent computing experience that full on 10" tablets can provide. Given that both smartphones and tablets use identical hardware, and overlap a great deal in functionality, its seems inevitable that they must eventually meet somewhere in the middle. I'd wager something between 5 and 6 inches might be the future. Alternatively, ideas are being tried to plug smartphones into docking stations resembling netbooks, with the possibility of connecting them to full-sized desktop monitors and peripherals some time down the road.
Certainly there will be a lot of companies engaging in some heated competition in this new market, moreso than any other computing field in recent memory. Cell phone makers in particular are entering the market, understandably so given the similarities between tablets and the field they already know. In addition to that, new players like PixelQi, Pandigital, and ViewSonic who only made Atom PCs before, are rising to the call of a super ergonomic computer. Such competition will evolve the new form factor probably faster than any other type of device. It's the wild west of computing again, like in the 80s when standards were being written and everyone was vying for authorship of it. We're not even bound by x86 anymore, and for the first time in decades the fundamentals of computing are being contested, and ARM is poised to lead that revolution. It's exciting to be a technology enthusiast again, and if you're not feeling it too, well then there must be something wrong with you.
