Could ARM's Cortex A9 Defeat x86 CPUs?

During our interview with Eric Schorn, VP Marketing for ARM's processor division, prior to the official release of the Cortex A9 MPCore, it was clear that the target of the nimble company was the lucrative multi-billion x86 segment where Intel rules.

ARM operates quite differently compared to Intel. Rather than owning the whole ecosystem, it only delves in selling intellectual property, essentially the building blocks, leaving its 200 partners to deal with the rest of the food chain including getting the samples and churning out the chips themselves.

Intel is still the biggest semiconductor supplier based on iSuppli figures with 2008 revenues of $33.7 billion. But then ARM has struck strategic partnerships with 10 of the other top 14 semiconductors firms worldwide which gives you an idea of the company's clout.

This allows more creativity, more flexibility and ultimately a more vibrant environment that supports a dozen operating systems and a multitude of derived ARM-based chipsets; a thriving competitive market with many niches.

This approach has helped ARM selling more than four billion cores in 2008, more than Intel has shipped in its entire lifetime and is reminiscent of what Linux did with the OS market (except that it is essentially open source).

But with the Cortex A9 MP core, ARM is looking even further; one thing that history taught us is that it is easier to scale up than scale down (eg: Nvidia vs ATI approach) and this is exactly what ARM is doing with the Cortex A9 MP.

The hard macro provided to its partners includes the possibility of scaling the processors to four cores officially. But I've been told that some vendors are looking even further with octo and 16-core models potentially coming up, depending on market demands.

Performancewise, the new Cortex-A9 is certainly no slouch with an estimated Coremark score of around 10,500 (for the 2GHz model) which makes it more powerful than a 3GHz Intel Xeon 5160 while consuming a fraction of the power.

Even the power optimised version which runs at 800MHz is as fast as an Intel Core 2 Duo clocked at 1.2GHz with a Coremark score approaching 4000 points. Given that the A9 can be scaled speedwise even further, the next generation Cortex could be even more powerful.

Although the chips are not compatible with desktop based Windows, nothing would prevent one of ARM's partners to come up with a hardware-based virtualised environment like Wine or using a solution like Win4Lin.

Interestingly, Microsoft is bound to release Windows 7 Mobile next year, just in time for the next generation Cortex A9 MP Core to debut on future smartphones. Amongst the other rival operating systems, Chrome OS, is set to be the more interesting of the bunch.

But there are a few issues that ARM needs to iron out. First, it needs to embark on a campaign to inform consumers (rather than B2B) about what it is, something similar to the Centrino or Intel Inside campaign. This can be done by enrolling its partners and creating a marketing fund.

Then there is the problem of delivery. Although many samples of non-mobile Cortex A8 have been announced, there are yet to reach a tipping point where they will flood the market a la Netbook. The problem is that Intel has the human resource to do just that, whereas ARM, which has only 1700 employees doesn't have the bandwidth to do that right now.