Simulation of integrated circuits consists of solving matrix-based equations. As the size of the modern circuits increases, the computation time and resources for a simulation have significantly increased. The recent progress in heterogeneous hardware platforms has created an opportunity to increase the efficiency of these simulations. This project demonstrates:

• The acceleration of lower-upper (LU) decomposition as the core algorithm in solving circuits using SYCL* and oneAPI on CPUs and GPUs
• Results and discussion of the development experience and future opportunities in using SYCL and oneAPI

Danial Chitnis is a Chancellor's Fellow and associate professor in electronics at the University of Edinburgh, Scotland. His research interest is modeling electronic and photonic systems.

Finlay Marno completed his master of science degree at the University of Edinburgh during which he developed his first GPGPU and SYCL program.