SALT-FM: Source Analysis Toolkit for Fortran and Mixed-Language Software

SALT-FM, the Source Analysis Toolkit for Fortran and Mixed-Language Software, will enhance developers' productivity and boost the efficiency of software applications written in various programming languages, including Fortran. Fortran finds extensive use in both private and government sectors, spanning computational fluid dynamics, chemistry, earth science, and weather prediction, often in tandem with other languages. Despite its prevalence, accessible and platform-agnostic tools for comprehensive performance analysis, static analysis, and program transformation remain scarce. Whether modernizing legacy code, porting components to languages like C++, or optimizing for contemporary hybrid computing systems and accelerators, few cross-platform solutions adequately support Fortran and its multiple recent language standards. SALT-FM will harness and leverage the government funded LLVM Flang project's robust capabilities, to develop a static analysis and program transformation toolkit. The funding will solely cover software development and requirement discovery. In Phase I, the focus lies on creating a prototype capable of instrumenting Fortran and mixed-language programs for profiling with the TAU Performance System. Subsequent efforts will prioritize refining and hardening the profiling and instrumentation tool and expanding static analysis and program transformation functionalities. These enhancements aim to facilitate tasks like language porting, interface generation, and legacy application modernization. Anticipated markets for SALT-FM include various industries such as aerospace, automotive, nuclear power, energy, manufacturing, earth science, weather prediction, chemistry, and life sciences, in addition to government entities like NASA, DOE, DOD, NRC, and others.