updated 05:55 pm EST, Mon February 6, 2012
Apple gets closer to using macroscalar name
Apple late last week quietly filed for a US trademark that shows it possibly getting closer to having its own chip architecture. After having slowly filed trademarks in countries outside the US, such as Trinidad and Tobago, the joint USPTO and Hong Kong application caught by Patently Apple suggests that Apple has been continuing development on a project it could be willing to show soon. Mentions of work had started as early as 2004.
The trademark itself owuld cover a technology that, bears some trace similarities to Intel's Hyperthreading in finding ways to keep a processor running closer to its potential. As a processor compiled code, it would make additional instructions that would keep a given code pipeline on a chipset perpetually occupied, with new code arriving just as soon as the old expires. Any resulting processor would not only be faster but would be more power efficient by getting tasks done in a shorter amount of time.
As Apple has to use largely off-the-shelf Intel processors in Macs, the macroscalar design if used at all would most likely apply to iOS devices. Although work began years before, Apple's acquisition of PA Semi gives it a wing that can customize processors to its own ends. The A4 and A5 have often been credited for giving Apple a speed advantage over components from NVIDIA, TI, and others.
Macroscalar wouldn't necessarily land in 2012-era iOS devices, if at all, but its more public appearance shows that Apple has at least enough interest to air details in the open. Most anticipation around this year's iPad and iPhone hinges either on a faster dual-core processor or a quad-core part.