Nvidia is really pushing CUDA parallel computing architecture along these days so Portland's Group announcement regarding their development of a compiler that will enable programmers to build and optimize CUDA applications to run on x86-based workstations, fits right in there with Nvidia's plans.

Developed in order to offload computationally intensive kernels to Nvidia GPUs, that are much better suited to perform highly parallel tasks, CUDA gives programmers explicit control over the mapping of general-purpose computational kernels to Nvidia GPUs, as well as the placement and movement of data between an x86 processor and the GPU.

This does provide developers with an API that makes harnessing the power inside the GPU a lot easier leading to the development of a series of applications that take advantage of the huge parallel processing power inside a GPU.

This does however need an Nvidia GPU in order to function, but PGI is developing a compiler for x86 platforms that will allow developers using CUDA to compile and optimize CUDA applications to run on x86-based workstations, servers and clusters with or without an Nvidia GPU accelerator, using the multiple cores and the streaming SIMD capabilities of Intel and AMD CPUs for parallel execution.

"CUDA C for x86 is a perfect complement to CUDA Fortran and PGI's optimizing parallel Fortran and C compilers for multi-core x86," said Douglas Miles, director, The Portland Group.  

"It's another important element in our on-going strategy of providing HPC programmers with development tools that give PGI users a full range of options for optimizing compute-intensive applications, while allowing them to leverage the latest technical innovations from AMD, Intel and Nvidia."

Although this sounds like an interesting innovation, I see this more as a way of tying developers to CUDA in some sort of plan to get their technology more widespread then Microsoft's DirectCompute or OpenCL, both of these APIs working on both AMD (formerly ATI) and Nvidia GPUs.

Of course that Nvidia would do whatever it can in order for this not to be the case and the compiler is meant to make more developers use the CUDA API when programming new applications that would see huge speed increases when an Nvidia GPU is used inside the system.