Apple recently made major changes to their laptop line, with several major updates from their previous iteration. Some of the more notable updates to me include the lack of FireWire in the MacBooks, the unibody Aluminum construction in both machines (which prompted a major board redesign), faster memory specifications, and the usage of NVIDIA’s platform in lieu of Intel.
You may say, “but the old ones had nVidia chips in them as well,” which is completely true, but it’s my guess that Snow Leopard (SL), Apple’s next generation operating system, will take the idea of general purpose graphics processing unit (GPGPU) and integrate it in two important ways. First, SL will take advantage of GPGPU on many OS level tasks that will improve performance. Second and perhaps more overlooked, I think that SL will ship with major new libraries in XCode 4, which will enable GPGPU integration in a typically programmer-friendly Apple way. In fact, it will likely be one of Apple’s core technologies, as mentioned on their site, especially if they write parallelization libraries that take advantage of the 16/32 stream parallel cores of the GPU. These innovations should benefit several machines released within the last couple of years, but these most recent updates represent a clear commitment to the cause.
Of course, more cores aren’t at all linearly scaling solutions, but there certainly are performance gains to be garnered from certain operations. I think GPGPU is brilliant right now for a few reasons. It turns out that quad core and higher in laptops will be a viable solution soon, but Apple demands portability and reasonable cooling on their laptops, so this is not quite ready.
GPGPU represents more than simply a stop-gap solution for Apple, while awaiting the solution to this quad core problem, because all of its hardware seems to be moving toward using NVIDIA graphics hardware, including their most recent MacBook and MacBook Pro updates.