Implementation of an FPGA-Oriented Complex Number Computation Library Using Intel OneAPI DPC++

OneAPI is introduced to increase the efficiency of application development work-flow by providing the same languages, programming models, debugging, profiling and optimization tools across CPUs and accelerators. In this paper, we implement a complex number computation library for FPGAs. The proposed library contains vector operations that can be executed in parallel on FPGA to increase the processing speed by 6.2 times. We use “operator overloading” in C++ to define all mathematical operation. The proposed library allows developers to write easy-to-read codes that can be developed and debugged on CPUs, while executed in parallel on FPGAs. Moreover, we implement the proposed library in Intel DevCloud and shows that we can easily use and distribute applications to cloud platforms.