The CUDA 6 parallel computing platform and programming model is designed to make programming easier, enabling software developers to dramatically decrease the time and effort required to accelerate scientific, engineering, enterprise and other applications with GPUs. Developers can accelerate applications up to 8X by replacing existing CPU-based libraries. Key features include Unified Memory, Drop-in Libraries and Multi-GPU Scaling, which can be used with the new BLAS drop-in library. The platform offers a full suite of programming tools, GPU-accelerated math libraries, documentation and programming guides.