Designing and Optimizing LQCD codes using OpenACC

An increasing number of massively parallel machines adopt heterogeneous node architectures combining traditional multicore CPUs with energy-efficient and fast accelerators. Programming heterogeneous systems can be cumbersome and designing efficient codes often becomes a hard task. The lack of standard programming frameworks for accelerator based machines makes it even more complex; in fact, in most cases satisfactory performance implies rewriting the code, usually written in C or C++, using proprietary programming languages such as CUDA. OpenACC offers a different approach based on directives. Porting applications to run on hybrid architectures “only” requires to annotate existing codes with specific “pragma” instructions, that identify functions to be executed on accelerators, and instruct the compiler on how to structure and generate code for specific target device. In this talk we present our experience in designing and optimizing a LQCD code targeted for multi-GPU cluster machines, giving details of its implementation and presenting preliminary results.