SIAM PP20 Minisymposium: Improving Productivity and Sustainability for Parallel Computing Software
As we move toward exascale machines and beyond, we face a daunting task of providing software for emerging extreme-scale disruptive architectures that will allow applications to run with extreme performance, high scalability, and reliability. In addition, demand for crosscutting software that spans multiple domain-specific sciences, a clamor for greater reproducibility, and new opportunities for greatly improved simulation capabilities, especially through coupling of data and physics across scales, will result in software being extremely complex and challenging. While hardware remains a primary focus for next-generation exascale computing, the aforementioned challenges also demand large investments in scientific software development.
In this minisymposium, we aim to cover lost ground and address growing technical, practical, and social challenges in software productivity, quality, and sustainability, with a goal of helping develop better and long-living, cutting-edge parallel computing software that can fulfill its designated roles in emerging exascale ecosystems. Along with sharing of diverse experiences by experts in the parallel computing field, the hope is that this minisymposium will motivate the community to realign and refocus on developer productivity and software sustainability---which are both urgent and essential---to build better scientific software.