x86 Everywhere: Two Years Later

In my original Atom architecture article I spoke about the benefits of having a platform that could run existing applications, in this case x86 applications. Developers don’t like porting to new hardware, which is one reason GPU computing hasn’t really taken off yet.

Since then we’ve seen a major change: the introduction of platform specific App stores. Starting with the iPhone App store and extending to most smartphone platforms (Android Marketplace, Palm App Store), with a simple way to sell their apps we’ve seen a completely new group of developers emerge specifically targeting smartphones. These aren’t your traditional developers. Companies like Adobe and Microsoft are effectively absent from any of the app stores. Instead what you find are smaller development houses putting forward smaller but very useful applications and games for use on these smartphones.

The scariest part for Intel is that none of these apps run on x86 hardware. While there are still more x86 applications than iPhone or Android apps, there are more smartphone friendly apps running on ARM architectures than x86. The advantage of being able to run existing code without lengthy port times just isn’t an advantage today. In fact, you could consider the move to x86 a disadvantage from the perspective of a company like Apple or Google. While it’d be simple to offer x86 versions of apps through a closed store system, it means extra work for the developer and for Apple with little benefit today. By aiming at the netbook first, Intel may have squandered one of its major potential advantages in the smartphone.

All isn’t lost however. There’s still the argument that the applications and algorithms that have yet to be moved to smartphones still exist in x86 form. As smartphones grow more powerful, so will the types of things we try to do on them.

ncG1vNJzZmivp6x7orrAp5utnZOde6S7zGiqoaenZIB3hZVooKeslaF6trrVnqClq12ivLC%2BxKyrqK%2BeYq6vsIytn55lkam8rnnZb2dpZaOav6qx0maroZ1dm660wMSsq2arnZa%2FtbzHqKWeZaCnvKSx0qymq2dn