Version 5.5 is the latest and greatest iteration of NVIDIA CUDA, NVIDIA's parallel computing platform and programming model that harnesses the power of GPUs. CUDA 5.5, sayeth NVIDIA, provides programmers with a robust, easy-to-use platform to develop advanced science, engineering, mobile and HPC applications on x86 CPU-based systems, and now on ARM-based ones as well. Beyond native support for ARM platforms, CUDA 5.5 delivers a number of new advanced performance and productivity features, including enhanced Hyper-Q support, MPI workload prioritization, and new guided, step-by-step performance analysis and fast cross-compile on x86. The latter feature enables developers to compile ARM code on fast x86 processors and transfer the compiled application to ARM. The v5.5 release also offers a full suite of programming tools, GPU-accelerated math libraries and documentation for both x86- and ARM-based platforms.
