Together with @rok_cesnovar and @Erik_Strumbelj we published a paper on OpenCL kernel generator. It describes how we automatically generate OpenCL kernel code from operations on matrices in C++ in Stan Math.
Paper link: https://dl.acm.org/doi/abs/10.1145/3388333.3388654
Link to video presentation (The last one in " Program of Talks"): https://www.iwocl.org/iwocl-2020/conference-program/#talks