Summary: This content delves into compiling Fortran code for WebAssembly to run in web browsers. Fortran, a decades-old language, still finds use in scientific and engineering applications due to its efficiency. Various methods, like Dragonegg, LFortran, and LLVM Flang, are discussed for compiling Fortran to WebAssembly, each with its drawbacks. The process involves modifying LLVM sources and techniques to bridge the gap between Fortran and WebAssembly. The post also touches upon calling Fortran subroutines from C, showcasing the complexity and challenges involved in the process. The goal is to enable running machine learning algorithms using efficient BLAS and LAPACK routines directly in web browsers.
https://gws.phd/posts/fortran_wasm/