Intel's Search For Relevance On Display At IDF

Intel recognizes we're in a PC-plus era--and just in time. But can it corral device partners who have moved on to other mobile hardware platforms?

Kurt Marko, Contributing Editor

September 13, 2012

5 Min Read

"Hey, remember us? Intel Inside? Blue Man Group? Dancing bunny suits boogieing down in the fab?"

You could almost hear the plaintive cry for relevance at the Intel Developer Forum (IDF) this week. Yet in a cruel twist of scheduling fate, the company's big confab that serves as part executive-level vision statement and part engineer-led technical deep dive had the misfortune of going up against the mother of all product rollouts, an event guaranteed to suck the oxygen out of the IT mediascape: an Apple iPhone launch.

Part of the problem is the stark contrast between Apple's clarity of purpose--producing the next generation of sleek, must-have mobile gadgets--and Intel's need to cater to such diverse constituencies. For Exhibit A, just look at the keynotes. The show kicks off with executive VP and chief product officer Dadi Perlmutter talking about "Reinventing Computing--Collaborating to Shape the Future From Datacenter to Devices." Perlmutter tries to make the tenuous connection between Intel's cloud ("the datacenter runs on Intel") and client bona fides, but he ends up spending the bulk of his time on how Intel is leading its OEM customers—sometimes, it seems, kicking and screaming--to out-Apple Apple. Of course, there was plenty of talk about Ultrabooks (aka MacBook Air knockoffs), "convertibles" (which looks like an iPad glued onto the lid of a MacBook Air), tablets (you-know-who again), and smartphones (where Cupertino just moved the goal posts)--all made better with Intel inside.

[ For more on Intel's vision for the future, see Intel Showcases New Chips, Big Vision. ]

Yet on Day Two we see Renee James, GM of the software and services group, focus on "Security and Services in an Age of Transparent Computing," a talk heavy on the developer ecosystem, application security, and the wonders of cross-platform language and instruction set compatibility--in other words, less vision and more about what's under the hood for developers. And in a sad bit of irony, James spends several slides extolling the merits of HTML5 as the technology enabling "transparent computing" on everything from phones to desktops. I guess she didn't get the news that just the day before and across town, Mark Zuckerberg said Facebook's "biggest mistake period was the focus on HTML5."

While Perlmutter was preaching the virtues of ubiquitous touch and speech interfaces, it's telling that Intel still recognizes the company's key advantages are its scalable CPU architecture, consistent instruction set, semiconductor process leadership, and manufacturing prowess. And with these, Intel still has a compelling story to tell.

Sticking to its Tick-Tock CPU release cycle, alternating between process shrinks and new processor architectures, the star of this year's IDF is Haswell, Intel's fourth-generation Core microarchitecture design. Leveraging last year's Ivy Bridge 22-nm, tri-gate fab process, Haswell is focused on reducing power without compromising performance. It achieves this by introducing several new power states, including an "extremely low-power active state" that Intel claims is a 20x (not 20%!) improvement from Ivy Bridge devices.

Haswell continues making slow but steady improvements to Intel's HD embedded graphics performance, but perhaps most importantly, introduces a modular architecture to the graphics engine that allows the same basic design to scale across devices with dramatically different performance needs and power budgets. Much like Intel has long done in stratifying its CPU products by core count, clock speed, and hyperthread support, the Haswell microarchitecture allows adding GPU "slices" to the same CPU design, effectively doubling performance of many graphics operations.

But Intel's low-power story doesn't end there. Where Haswell--which Intel claims can scale down to 10W TDP, a level still too power-hungry for tablets and smartphones--leaves off, Atom picks up. Long treated like Intel's ugly stepchild, a significantly refreshed Atom lineup is tailor-made for handheld devices with their small batteries and always-on availability. Sadly, Atom still must live with the older-generation 32-nm fab process, but its SoC design, including everything from CPU core (or cores) and L2 cache to graphics engine, video codec, and I/O modules, is squarely aimed at devices currently using Nvidia's Tegra or Qualcomm's Snapdragon platform--all while maintaining 32-bit x86 compatibility. Intel claims the base, single-core (with hyperthreading) Medfield part uses only 14mW in standby and 850mW while browsing, but if you have the power budget, you can step up to the dual-core Clover Trail part for double the performance.

At IDF, Intel highlighted its low-power mastery with its application to mobile devices, but the technology is equally applicable to the data center, where increasing server density, and accompanying demands for more power and cooling per square foot, are rendering many facilities obsolete for cloud-era workloads.

The client focus is natural, since Intel historically first introduces new architectures on parts designed for laptops, not servers, and Haswell will hew to tradition. But it will be interesting to see how the technologies on display this week manifest themselves in the Xeon (Core CPU) and Centeron (Atom) lines. We're especially interested in how Intel will react to ARM's pending release of a 64-bit architecture, which could debut in prototype systems later this year, potentially shaking up the budding microserver market. It's a topic we'll explore in more depth in this year's InformationWeek State of Servers 2012 report in a few weeks.

Make no mistake: Intel is still a technology powerhouse, leading the industry on several fronts, from CPU power and performance to high-speed network I/O and embedded hardware security, and the fruits of its engineering might were on full display this week at IDF. Although it risked missing the mobile migration by protecting its core CPU business for too long, it's clearly had an epiphany. If not fully embracing the post-PC era, it recognizes we're in a PC-plus era--and just in time.

Now the challenge is corralling OEM partners that have already moved on to other mobile hardware platforms. Although Microsoft is largely on board, Intel must still close the deal with the device manufacturers. It's not too late, but as we learned over the last couple weeks, Apple, Amazon, and Google aren't standing still.

Find out the nine questions you must ask before migrating apps to the public cloud in the Cloud Ready? special issue of InformationWeek. Also in this issue: It's time to lay to rest two common myths of the cloud computing era. (Free registration required.)

Read more about:

20122012

About the Author(s)

Kurt Marko

Contributing Editor

Kurt Marko is an InformationWeek and Network Computing contributor and IT industry veteran, pursuing his passion for communications after a varied career that has spanned virtually the entire high-tech food chain from chips to systems. Upon graduating from Stanford University with a BS and MS in Electrical Engineering, Kurt spent several years as a semiconductor device physicist, doing process design, modeling and testing. He then joined AT&T Bell Laboratories as a memory chip designer and CAD and simulation developer.Moving to Hewlett-Packard, Kurt started in the laser printer R&D lab doing electrophotography development, for which he earned a patent, but his love of computers eventually led him to join HP’s nascent technical IT group. He spent 15 years as an IT engineer and was a lead architect for several enterprisewide infrastructure projects at HP, including the Windows domain infrastructure, remote access service, Exchange e-mail infrastructure and managed Web services.

Never Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.

You May Also Like


More Insights