A very insightful article about the implication of this LLVM Intermediate Representation format on the future CPU in the Apple Watch, or in the Macintosh:
LLVM turns an app’s source code into Bitcode, and then turns that Bitcode into an executable app. This design makes it incredibly simple to add support for new languages (front ends), and for new CPUs (back ends). While Bitcode itself can’t run on anything, it can be transformed into any supported CPU architecture, including ones that didn’t exist when the app was submitted.
UPDATE: There is a thread at yCombinator from drfuchs that has some clarifications on what is feasible with Bitcode:
I managed to ask Chris Lattner this very question at WWDC (during a moment when he wasn’t surrounded by adoring crowds). “So, you’re signaling a new CPU architecture?” But, “No; think more along the lines of ‘adding a new multiply instruction’. By the time you’re in Bitcode, you’re already fairly architecture-specific” says he.
Wednesday, June 17, 2015
Copyright © 2015-2018 Selected Links | RSS | Twitter | Linked list