Intel and NVidia Team Up in Parallel Programming Initiative

While Intel and NVidia have respectively claimed that GPU are dead, and CPU have no future, they found themselves common interest in an exciting project which could, of course, serve their goals: the Parallel Programming Initiative
Indeed, even if future computer will only run with a GPU, or a CPU, having multi core processing units will not bring that much extra performance gain if software (and OSes) are not able to benefit from such raw power, in order to deliver massively multithreaded applications.
Of course, Intel and NVidia are not the only sponsors of this initiative, and AMD, Nvidia, Sun Microsystems as well as Intel, Hewlett-Packard and IBM, are also on the list. By providing hardware and financial support to university laboratories, such as Stanford’s Pervasive Parallelism Lab (PPL), those sponsors expect to boost research in developing parallel programming. However, all companies might well be willing to push their own project, with Nvidia and its Tesla stream processor cards, AMD’s GPU, AMD’s and Intel’s CPU and Intel's massively parallel Larrabee accelerator card. Such basic research might also make developers' life easier by providing a platform for native multi core support, so coding application without thinking of core management.
All those companies have a common interest: current business plan for computer industry development is based on the multiplication of the number or cores per GPU/CPU, so if such raw power can not be used, performance gain will never be really noticeable by end users.