xrt VS pyopencl

I am working on simulations in the x-ray range, so fundamentals like propagators are the same, but approximations of optical elements and interactions are different. The x-rays are emitted by a synchrotron radiation source, i.e. an electron storage ring. The statistical properties of the wavefronts are determined by the phase space of the electrons in the storage ring, which is usually modeled by a 5D Gaussian distribution (space + momentum + energy). People are interested in the coherence properties of the emitted ensemble of wavefronts, so the simulation consists are drawing random parameters from the 5D Gaussian, calculating the wavefront emitted and propagating it to the points of interest. There, the wavefronts are collected and their coherence properties are evaluated. I use the coherence module of XRT for that: https://github.com/kklmn/xrt/blob/master/xrt/backends/raycing/coherence.py

Please note that OpenCL consists of two parts: host API and a separate language which is used to write kernels (code which is going to be offloaded to devices). OpenCL specification describes host APIs as C-style APIs and that is what implementors has to provide. However, there are number of various libraries which provides bindings for other languages: - C++ - Python - Go - Rust

I thought the project could be dead, but then I looked into the latest commits to the repository, and it is certainly not dead as a project.