Accelerated Computing with Standard C++, Python, and Fortran

It has long been thought that accelerated computing required experience with specialized programming languages, directives, or language extensions, but as parallel programming has become more mainstream, standard programming languages have made parallel programming a feature native to the language. Programmers can now write parallel-first code, no longer adding parallelism as an afterthought, but as the baseline. I'll highlight recent advances in native language parallelism in C++, Python, and Fortran and show how programmers can begin to support GPU acceleration as a native feature in their applications with no specialized code.