In terms of hardware, the (parallel) computing power of graphic processing units (GPUs) might provide significant performance benefits to R users. Following on the earlier releases of vendor-specific APIs (e.g., nVidia (2007) proposed CUDA), standards for programming GPUs in a vendor-independent fashion are beginning to emerge (Khronos Group 2008).
They offer a programming model that is designed to allow direct access to the specific graphics
hardware, with the graphics hardware running a very high number of threads in parallel. A bioinformatics application for sequence alignment with GPUs (C code, no R integration) has been published by Manavski and Valle (2008) and illustrates the usability of GPUs for acceleration and management of large amounts of biological data.
They offer a programming model that is designed to allow direct access to the specific graphics
hardware, with the graphics hardware running a very high number of threads in parallel. A bioinformatics application for sequence alignment with GPUs (C code, no R integration) has been published by Manavski and Valle (2008) and illustrates the usability of GPUs for acceleration and management of large amounts of biological data.