We must ensure the reliability of scientific computing hardware and software systems that are fundamental to enabling discoveries through high performance computing (HPC) systems that push the edge of computing capabilities. Similarly, the reliability and dependability of systems that make data-driven decisions through large-scale machine learning (ML) must also be ensured, as these systems not only control day-to-day aspects of human life but also interact with HPC systems in significant ways to reduce and/or accelerate simulation.

Stakeholders in correctness in this space belong to several sub-disciplines of computer science: computer architecture researchers who design special-purpose hardware that offers high energy efficiencies; numerical algorithm designers who develop efficient computational schemes based on reduced precision as well as reduced data movement; and researchers in programming language and formal methods who seek methodologies for correct compilation and verification. We must pursue formal and informal correctness, resilience, reproducibility, trustworthiness and explainability of results.

Fortunately, the FCRC attendees are such stakeholders, making CSC an ideal fit within it. We invite you to register for CSC, ask questions, and contribute to the discussions.

Maya Gokhale (LLNL)

Ganesh Gopalakrishnan (University of Utah)

Jackson Mayo (Sandia National Labs)

Santosh Nagarakatte (Rutgers University)

Cindy Rubio-Gonzalez (University of California, Davis)

Stephen Siegel (University of Delaware)

Call for Participation

We invite you to register for CSC, ask questions, and contribute to the discussions.

By participating in CSC, you will hear keynotes from prominent speakers (about 2) who have worked on the aforesaid issues as well as speakers in several lightning talks (about 9) each of 5 minutes. But most importantly, you will be informing DOE and NSF through our report which will reflect your input (and will be open for further comments over the summer).