Intel’s New Mobile Chips – Still Not There Yet

January 7, 2013 - Written By Lucian Armasu

Intel has just announced 2 new mobile chips, one that is a quad core regular Atom chip, codenamed Bay Trail, that’s destined for tablets mainly, and is meant to compete with the likes of Tegra 4, Exynos 5 Quad and S4 Pro, and the other is one that is meant to compete more with Cortex A7 for the low-end markets. The quad core Atom chip is supposed arrive at the end of this year, while lower-end Atom Z2420 that is meant for low-end goes up to 1.2 Ghz, and is also supposed to arrive later this year.

So why is Intel still playing catch-up? Let’s take the quad core Atom for instance. Cortex A15 is now up to 65% faster than Atom, even when Atom is slighter higher clocked (1.8 Ghz vs 1.7 Ghz), and under the exact same OS (like Chrome OS, or Linux). Intel Atom has always had the lead with Atom compared to ARM chips when it comes to performance.

That’s because it was designed to be a more powerful chip (and therefore more inefficient) from the beginning. So Intel didn’t have to worry as much about performance ever the years, but they had to improve on energy efficiency, if they want to catch-up with ARM. While they seem to have pretty much gotten there, it also seems that they slipped from having the most powerful mobile processors.

As for their lower-end single Atom chip that’s meant for emerging markets, it seems even further behind compared to dual core or quad core Cortex A7’s that will be all over in the emerging markets, even in under $50 devices. Cortex A7 is so efficient and so small, that it only measures 0.5mm2 at 28nm. That is about 10x smaller compared to Cortex A8 at 45nm, and yet it’s about as powerful.

Another important factor that manufacturers are going to care about for both the quad core Atom and the single core low-end one, is price. Intel’s Atom chips are still significantly more expensive than their ARM equivalents (50%-100% more), which gives little incentive for manufacturers to prefer it over the well known ARM architecture that they’ve been using for years in smartphones.

In the end, this might be Intel’s biggest problem – not the fact that they are still behind from a performance and/or energy efficiency point of view – but that even if they fix those problems, and reach parity with ARM for both performance and efficiency (and GPU performance as well, which is where Intel is more behind than any single ARM chip makers right now), they will still need to offer competitive prices as well. And that might be the toughest thing Intel will have to do, because Intel is not a company that can survive on razor sharp chip profits, like the ARM chip makers.