Homepage Inria website

Section: Scientific Foundations

High-performance computing

High-performance computing is another tool that we will use to achieve our goals. It covers several paradigms: grids, single-instruction, multiple-data (SIMD) instructions, graphics cards (GPU). In a near future, processors may offer tens or hundreds of cores with large vector units, combining again several levels of parallelism. Libraries like CUDA and OpenCL also facilitate the use of these manycore processors. This new hardware architecture brings promising opportunities for time-consuming bottlenecks arising in bioinformatics.