Enabling RAJA on Intel GPUs with SYCL

To achieve high performance, modern HPC systems take advantage of heterogeneous GPU architectures. Often these GPUs are programmed using a vendor preferred parallel programming model. Unfortunately, this often results in application code that is not portable across vendors. To address this issue, open parallel programming models have been introduced. One such parallel programming model is provided by the RAJA Portability Suite. RAJA is a portability layer that provides an abstract application developer API as a library through modern C++. In RAJA, computational kernels are lowered down to a backend language appropriate for the target architecture. Additionally, RAJA is designed to provide control to the application developer over kernel execution while minimizing modification to the application source code.