Skip to content

Towards co-execution of massive data-parallel OpenCL kernels on CPU and Intel Xeon Phi

Heterogeneous systems composed by a CPU and a set of di erent hardware accelerators are very compelling thanks to their excellent performance and energy consumption
features. One of the most important problems of those systems is the workload distribution among their devices. This paper describes an extension of the Maat library to allow the co-execution of a data-parallel OpenCL kernel on a heterogeneous system composed by a CPU and an Intel Xeon Phi. Maat provides an abstract view of the heterogeneous system as well as a set of load balancing algorithms to squeeze the performance out of the node. Experimental results show that this approach always outperforms the baseline with only a Xeon Phi. Furthermore, the load balancing algorithm has a huge impact in the system performance, therefore, the right selection is essential.