Understanding and enabling Performance Portability is a key research area of the High-Performance Computing Research Group here in Bristol. We’ve published a number of articles on the topic, and have been using a metric, originally developed by Pennycook, Sewall and Lee at Intel, to quantify the performance portability of applications. The metric was first introduced at the PMBS Supercomputing Workshop in 2016, and uses an unusual symbol that looks like a double “PP” character to represent performance portability. This is a non-standard symbol that is constructed using some LaTeX commands, which are shared below with the permission of the original authors. Feel free to use this code to generate the symbol.

% Packages

% Command to use in Math Mode

% Command to use in text