Parallelising semantic checking in an IDE: A way toward improving profits and sustainability, while maintaining high-quality software development

Abstract

After recent improvements brought the incremental compilation of large industrial test suites down to a few seconds, the first semantic checking of a project became one of the longest-running processes. As multi-core systems are now the standard, we derived a parallelisation using software engineering laws to improve the performance of semantic checking.Our measurements show that even an outdated laptop is fast enough for daily use. The performance improvements came without performance regressions, and we can’t expect additional massive benefits even from infinitely scaling Cloud resources.Companies should utilise cheaper machines that still o er enough performance for longer. This approach can help businesses increase profits, reduce electronic waste and promote sustainability while maintaining high-quality software development practices.