The old “more cores” argument
I don’t buy into the “more cores are always better” argument, after all, Intel’s latest benchmarks for its dual-core chip easily compete with top of the line handsets like the Galaxy S4. But Intel has been very slow at adapting to the need/demand for multi-core chips in the mobile market. Smartphones have been using quad-core chips for more than a year, yet Intel’s newest processor, Clover Trail +, is still plodding along with two cores.
This isn’t to say that Intel’s processors are slow, but there are many reasons why multi-core chips are particularly good in the mobile market, and Intel has been so slow to capitalise on this.
For a start, multi-processing is quite important when you have a lot of background tasks going on. Then there’s the energy saving potential, where cores can be clocked lower or shut off whilst still maintaining some level of device performance. Finally, the most important argument is price. Sure, Atom is designed to be a low energy yet powerful CPU, but is rather inflexible when it comes to the price aspect of mobile devices.
Compared this to ARM: the cores are small but effective in low end devices, yet easy to stick together into low cost multi-core solutions. For example, the eight-core Exynos 5 Octa only cost Samsung $30 in the Galaxy S4. Chip for chip Intel may be faster now, but when it comes to performance per dollar ARM is out in front. It’s this price flexibility which allows ARM to simply sidestep Intel.
A quad-core mobile processor is certainly something that could help Intel finally break this market, but the four-core Bay Trail Atom CPU isn’t scheduled for release until Q3 2013, and then we still have to wait for utilizing devices to be made available too. On top of that, the price to performance ratio has to be right. ARM is still cheap and pretty nimble, Intel has a lot to improve on here.
Lack of LTE support, Intel needs a complete mobile SoC
Another big drawback for manufacturers is that Intel is yet to provide an all-in-one LTE compatible chip, which again makes Clover Trail a more expensive choice for high end handsets. This shows, as Intel’s only successes in the mobile market have been in tablets, where networking isn’t such a huge issue.
Even so, this is probably the number one problem with Intel’s chips. They lack a lot of standard features available in other smartphone chips. Nvidia realised that it had the same problem with the Tegra series, and decided to design the Tegra 4i with built-in LTE. We know that Intel is working on built-in LTE functionality, as the company plans to ship out data-only versions of its LTE modem integrated circuits first, followed by data and voice capable chips, sometime during 2013, but we’ll be waiting until 2014 before we see any compatible handsets. However, this should go some way to improve Intel’s market presence.
This also brings us back to the earlier point. ARM doesn’t have to worry about designing entire packages for mobile systems. Instead companies like Qualcomm can spot the demand, and adapt the design to fit. Whilst Intel has been focussed on bringing the performance and power consumption up to par, it has failed to keep up with the secondary features being offered by its competitors.
Deals with other manufacturers
So far Intel hasn’t managed to secure long-running contracts with smartphone or tablet manufacturers, but what if Intel was to team up with another big company? Well, again, Intel appears to be addressing my concerns. According to Forbes, Intel is supposedly in talks with “a potential large, unidentified mobile customer”, which could possibly be Apple. Since the recent dispute with Samsung over chips for its iPhone products, Apple has no doubt been looking for another company to fill the void, and diversifying away from being so reliant on its main smartphone competitor makes a lot of sense.
Intel already has a few other contracts regarding Windows laptops and Apple MacBook chips, but a deal for iPhone or iPad products would be a huge milestone, as the smartphone market is one which Intel has been struggling with. Sadly though for Intel, nothing is confirmed and, at least for the immediate future, Intel will have to make do with it’s current state of affairs.
Flexibility is the key
In the end, I’d say that Intel’s major problem has come from a lack of product flexibility. The mobile market encompasses a wide range of products, from high performance tablets, to energy efficient smartphones, and of course, budget products. Intel’s product roadmap so far has failed to meet the variety of this demand, and the future roadmap is far from convincing either. Bay Trail may help, but with LTE support still not scheduled until next year, Intel is still going to lag behind Samsung and Qualcomm.
In my opinion, if Intel wants to break into the smartphone and tablet markets there needs to be a fundamental shift in the way the company does business. Simply churning out a few chips, no matter how good they are, is unlikely to make a big enough impact. Look at Nvidia — a single Tegra chip every year doesn’t capture as much of the market as the broad range of chips available from Qualcomm. At least Nvidia has made the effort to develop the Tegra 4i, which is designed specifically with smartphones in mind.
However, Intel has taken a lot of steps recently which could put it in good stead for the coming years. Improving the quality of its chips is one part, and teaming up with other manufacturers will also be key to breaking this market. The next line-up of processors will certainly be Intel’s best attempt yet, but we’ll have to wait and see whether or not the company has done enough to interest smartphone and tablet manufacturers.