LAS VEGAS--While CES has previously been a show filled with big announcements and brand new technologies, this year’s show was a little different, with more incremental advancements centered on refining devices people already own.
There are a lot of things emerging that just make existing products work better and smarter,” said analyst Jack Gold, though he noted those smaller improvements could still make a huge difference to consumer experience.
Perhaps the one big product announcement at CES 2012 was Intel’s reference design smartphone, and subsequent deal to distribute an Atom powered device in China, but whether consumers will be sold on mobile with Intel inside remains a discussion point.
“Intel can do a lot of fine tuning, they have a lot of software compiler expertise, so it will be interesting to see how well it runs,” said Gold, noting that people would take note if Android ran better on the Atom platform than it currently runs on ARM, adding, “there is a chance that will happen.”
Gold explained that responsiveness would be the real key, saying “If I can open a browser in one second instead of three seconds, people will notice that. If I can stream media at 30 FPS instead of 10 frames per second, people will notice that. If I can run applications that can open up in microseconds instead of two or three seconds, people will notice that.”
Turning the discussion to the buzz around Windows 8, Gold expressed excitement for a more “conglomerate experience” which would see the OS becoming more unified between phones, tablets and PCs.
As for Windows 8’s chances on ARM, however, Gold was not optimistic.
“Doomed is maybe too strong a word but I don’t think it’s going to be terribly successful,” he posited. Gold’s skepticism arises from the lack of backwards software compatibility on the platform and said the danger is that if the market doesn’t quickly take up Windows on ARM, Microsoft will stop spending resources on it, as has been the case with other platforms beforehand.
Meanwhile, Gold explained how graphics cores were becoming ever more critical going forward, be they on smartphones, tablets or PCs, with most browser tech today being GPU driven.
I am not sure it's just a matter of changing their recipe and optimize for power. It ain't that simple. To do this, they have to depart from their core x86 belief, which they do not seem to want to do.
What about all the XP apps that failed to run in Vista and then Windows 7? Some people never left XP. Others bailed on Windows completely. But those who remained, those running Windows 7, about 1/3 the PC market, by self selection are not so attached to legacy code.
I think Windows is, and for some time has been, irrelevant. It's only market advantage is the stable of software and that runs only on Intel. Vendors will not rewrite anything but the most profitable apps (Adobe, MS Office) for ARM because it costs too much and is more a rewrite than a recompile. Look how long it took vendors to support iA64 (64 bit Windows). That was a trivial change compared to Intel vs Arm.
Maybe not. Another artice in this issue had a talking head from Intel who claimed their past preference for computing power over power efficiency was simply that, a choice. That makes sense given that, in the PC market, you could only sell your CPUs for $1000 each ("Extreme" version) if it was faster than everyone elses CPU.
Now they just need to change the recipe and optimize for power consumption. Intel's foundries are at least one geometry node ahead of the ARM guys (ie TSMC). Each node increases efficiency considerably. So they ought to be able to make a better performing part at equal or even lower power.
"it will be interesting to see how well it runs,” said Gold, ...people would take note if Android ran better on the Atom platform than it currently runs on ARM, adding, “there is a chance that will happen.”
What makes anyone believe that it will perform 'noticebly' better? Yes there is a 'chance' but more likely it won't be noticeable.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.