Linear Algebra Libraries with DAG Runtimes on GPUs
Résumé
Nowadays many clusters integrate GPUs accelerators in their architectures that provide a huge amount of computational units rarely fully exploited. We present in this talk how tile algorithms and DAG schedulers as PaRSEC or StarPU can allow the programmer to integrate GPUs in their algorithms. We will present dense linear algebra algorithms as Cholesky or LU factorizations that exploit distributed architectures equipped with GPUs.