high performance computing on graphics processing units: hgpu.org

The Large Hadron Collider (LHC) at CERN will see an upgraded hardware configuration which will bring a new era of physics data taking and related computational challenges. To this end, it is necessary to exploit the ever increasing variety of computational architectures, featuring GPUs from multiple vendors and new accelerators. Performance portable frameworks, like SYCL, allow to offload the computational work on non-CPU resources, while retaining their performance, without the need to maintain different implementations of the same code. The High Energy Physics (HEP) community employs a wide variety of algorithms and tools for accelerators, but it still lacks a streamlined coherent approach that can target many use cases without compromising the usability aspect. In this paper, we present our efforts in creating GenVectorX, a C++ package that provides classes and functionalities to represent and manipulate particle events using the SYCL programming model. The SYCL-based implementation exhibits comparable performance and scalability as the CUDA implementation when targeting NVIDIA GPUs.