At this year’s CES, both Intel and AMD confirmed that they’ll support dual-mode tablets going forward. The idea from both companies is that consumers will be able to buy a Windows tablet running Android. The two companies are pursuing this goal in somewhat different ways; AMD is publicly allying with BlueStacks (again) to offer an emulation environment for Android apps, while Intel is apparently offering native x86 execution for Android in both environments.
Parsing the differences
Further muddling the matter is the fact that BlueStacks has made statements implying that because the AMD APU contains an embedded ARM controller, AMD chips could outperform their Intel counterparts. Here, we need to set the record straight a bit.
The only embedded ARM core we’re aware of on any shipping AMD product to date, including Kaveri, is a Cortex-A5 “Security Processor” that the company expects to ship in upcoming tablet designs in 2014. The entire Trustzone concept was designed to provide a low-power security processor. It’s worlds away from a high-end integrated ARM core, and even if it wasn’t, there’s only one of them baked into next-gen APUs.
A single Cortex-A5 isn’t going to give AMD some overwhelming performance advantage given that Android can already run natively on x86 to start with. Intel and AMD are chasing two different visions here – Intel claims you’ll be able to switch between different operating systems at the touch of a button, while AMD is talking up the ability to launch apps from the Windows 8.1 desktop or Start screen. Intel, in other words, is talking up the option of having Android as a discrete operating environment (presumably virtualised within Windows). AMD is talking about an enhanced option to run Android applications from within Windows.
Is the mass market going to care?
The ability to run Android applications from within or alongside Windows is an obvious benefit to anyone who already finds themselves juggling multiple devices. That mostly means developers and IT personnel, and it’s not a large group of people in absolute terms. Outside of the tech community, most people buy devices based on what they’re familiar with and what applications the product can already run. Windows 8 has struggled to find a footing for itself precisely because Metro mucked about with much of what people expected from a PC, while the App Store has failed to deliver application-level parity with its Android or iOS counterparts. Windows 8 already presents two user environments (poorly); tossing Android in as a third seems ill-advised.
The AMD-BlueStacks approach of emulating applications rather than launching the full operating system is an easier way to offer the advantage of Android compatibility without the drawbacks, but it comes with its own set of challenges. AMD’s share of the touchscreen-equipped market is limited and Android applications are themselves built to different UI specifications than their Windows counterparts. Most are touch-centric, and the keyboard-and-mouse are an imperfect substitute (despite all of Microsoft’s claims to the contrary).
The only way for Intel or AMD to really pull this concept off is to offer a definitive best-in-class experience that can match anything an Android-only product can produce. Apple’s share of the computer market rose after it began using Intel chips partly because the machines themselves were more competitive, but also because – thanks to Boot Camp – Apple products were competitive with Windows PCs when running the Windows operating system. It’s not enough to offer a middling level of compatibility – any system with multi-OS capability has to compete equally across both environments.
If Intel is virtualising Android within Windows, it’s going to have trouble putting a perfect polish on the experience without sacrificing battery life or execution speed. AMD’s approach may make more sense for users who only care about a handful of games or similar utilities, but AMD is ultimately dependent on a third-party’s software framework and product support. The ARM core integrated in upcoming APUs may be of some limited assistance, but it’s not going to fundamentally change the performance game. Nothing short of an embedded ARM block would make that happen, and at that point, you might as well just reboot into Android natively.
Google and Microsoft aren’t backing this endeavour, but I’m not sure their support would materially alter the outcome. Dual OS may have a compelling use case for some, but it’s unlikely to find more than niche acceptance. Most Android users already own other Windows devices, while most Windows owners bought into the Windows tablet ecosystem precisely because they needed Microsoft’s feature set. A few cross-platform apps offered by emulation may be icing on the cake, but not much more.