TuCCompi: A Multi-Layer Programing Model for Heterogeneous Systems with Auto-Tuning Capabilities

During the last decade, parallel processor architectures have be-come a powerful tool to deal with massively-parallel problems that require High Performance Computing (HPC). The last trend of HPC is the use of heterogeneous environments, that combine dif-ferent computational power units, such as CPU-cores and GPUs. Performance maximization of any GPU parallel implementation of an algorithm requires an in-depth knowledge about its under-lying architecture, becoming a tedious task only suited for ex-perienced programmers. In this paper we present TuCCompi, a multi-layer framework that not only transparently exploits hetero-geneous systems, but automatically tunes the GPU capabilities by choosing the optimal values for their configuration parameters, us-ing the kernel characterization provided by the programmer. This model is very useful to tackle problems characterized by indepen-dent, high computational-load tasks with none or few communi-cations, such as embarrassingly-parallel problems. We have eval-uated TuCCompi in different, real-world heterogeneous environ-ments using the APSP problem as a case study