I'm Dr. Moritz Lehmann, theoretical physicist at the University of Bayreuth, biophysics graduate at Elite Network of Bavaria, and member of DLR Graduate Program and Khronos OpenCL Advisory Panel. I'm one of the world's leading experts in high performance computing on GPUs using OpenCL/C++.
My main project right now is FluidX3D, the fastest and most memory efficient lattice Boltzmann CFD software, running on all GPUs via OpenCL. It comes with integrated interactive raytracing graphics that I implemented in OpenCL entirely from scratch.
I know all of the techniques to optimize software to the (roofline model) limit. I like extremely efficient, short and aesthetic code without any bloat or overhead, and I'm not a fan of external libraries as they often lead to compatibility issues and elongate compile time disproportunately. Instead I come up with my own much faster and more elegant solutions, for example math/string utility functions, FP32<->FP16 floating-point conversion or writing an integer array to a .bmp image file.
Want to learn OpenCL? I created this OpenCL-Wrapper for the C++ bindings to make it a lot simpler while keeping functionality and performance.