A few weeks back Intel made news announcing its 3D transistors being commercialized. This new Tri-Gate technology can enable increased computing power at reduced energy levels. We all know that this sort of breakthrough is crucial in mobile computing and even general electronics can benefit from increased performance at lower power draws. Think about the benefit of such technology in data centers for example.
Over the years, specialized processors have had to compete with Intel’s CPUs which gave so much more computing power per dollar. Intel had the advantage because it was able to produce so many processors that the price for each dropped dramatically and they passed along some of these savings to customers. But this meant that specialized hardware in some cases was replaced by software running on Intel CPUs. Case in point is the DSP resource board market where companies like Dialogic, NMS and Brooktrout were leaders in the nineties – powering voicemail, IVR and complex government speech detection scouring phone networks for keywords which could compromise national security.
In the nineties in fact, Intel picked up market leader Dialogic so it could migrate its board-based solutions to HMP which stands for host media processing. In this manner the world’s leading processor company decided it could rapidly evolve and even more directly benefit from the move to speech processing via its processors.
It turns out the transition took place but the communications market was different enough from the processor market that Intel eventually sold Dialogic.
We know smartphones, tablets and other handheld gadgets are a huge and rapidly-growing market and ARM is king in this realm. Interestingly, Intel too can produce chips based on the ARM-architecture but the margins are insufficient as you might imagine because they are forced to compete on an open playing field where their massive overhead is a disadvantage.
So the company instead is going after the mobile market with its Medfield processor which is meant to take on ARM chips more effectively with better power/performance characteristics in some instances such as playing video. But the company is far behind as quad-core ARM chips are around the corner while Medfield-based chips are first available in the single-core variety. If that isn’t a big enough problem, the company is facing delays with this chip as well – as reported by TMCnet earlier today.
By 2014, Intel’s 22 and 14 nm technology is said to be coming to its processor line – this is the technology which was discussed as part of the 3D processor announcement. But the question the market will ask is what sort of advantage will Intel need to offer to persuade hardware vendors to change sides? Moreover, what will be the advantages of switching to Intel designs? Of course Intel could have lots of tools up its sleeve which will make programming for these chips easier – making it similar to a PC but by 2014, the state-of-the art in ARM technology will have advanced greatly as well.
One has to wonder what sort of leap Intel needs to make to convince the world of smartphone and tablet computing to change sides. And it isn’t conceivable to think the mobile market in-part anyway may be beyond the company’s reach.