Numerical library reuse in parallel and distributed platforms

International audienceIn the context of parallel and distributed computation, the currently existing numerical libraries do not allow code reuse. Besides, they are not able to exploit the multi-level parallelism offered by many numerical methods. A few linear algebra numerical libraries make use of object oriented approach allowing modularity and extensibility. Nevertheless, those which offer modularity together with sequential and parallel code reuse are almost non-existent. We analyze the lacks in existing libraries and propose a design model based on a component approach and the strict separation between computation operations, data definition and communication control of applications. We present then an implementation of this design using YML scientific workflow environment jointly with the object oriented LAKe (Linear Algebra Kernel) library. Some numerical experiments on GRID5000 platform validate our approach and show its efficiency