The dual GPU system implemented in Apple's new MacBook Pro notebooks is not Nvidia's Optimus technology, it has emerged.
Instead, Apple uses multiplexer (mux) hardware to switch the output to the screen between the processor-integrated graphics core and the discrete Nvidia GPU, we hear.
Like Optimus, the switch is triggered in software by Mac OS X which monitors when frameworks that require GPU acceleration come in to play. At that point, it powers down the IGP and powers up the GPU.
We haven't seen the new MacBook Pros yet, so we can't say if Apple has eliminated the usual screen flicker that takes place when one graphics core's output halts and another's begins to be fed to the display through the mux.
Optimus is a more elegant solution. It abolishes the need - and cost - of mux hardware by feeding the GPU output straight into the IGP's frame buffer. That means it can be switched in at will without screen flicker. The switch takes place automatically according to need, as measured by Nvidia's middleware, which monitors which applications the user is running.
The effect is the same: increased battery life because the IGP is running more of the time than the GPU is. Ars Technica, which spoke to Apple about the technology, notes that, unlike Optimus, Mac OS X powers down the IGP when it's not in use - Optimus can't do this because it uses the IGP as a display controller - and that this boosts battery life.
We'd note, though, that the GPU's power drain is likely to outweigh any advantage from switching off the IGP. The real gain comes from turning off the discrete graphics whenever possible.
Ars does usefully reveal that Mac OS X's monitoring system is triggered by the use of OpenGL, Core Graphics, Quartz Composer and the like. These frameworks are also used by the OS simply to handle the UI, so it would be interesting to learn exactly how Apple distinguishes between software than needs the GPU - a game, say - and software whose graphics and compositing requirements can be satisfied by the IGP. ®