23 September 2021 – Assistant Professor Kuldeep Meel and his research collaborators, Dr Mate Soos and Nicholas Prevot, have won the Spring 2021 Amazon Research Award for their project titled ‘GPU-Enabled Parallel SAT Solving’, which focuses on enabling scalability of parallel SAT solvers using Graphics Processing Units (GPUs).
“GPUs have played a big role in scalability of modern machine learning systems, and we envision that they can play a similar role in the context of SAT solving. Our project seeks to make this possible by developing a shared library, GPUShareSAT, that can be easily plugged into standard SAT solvers,” said Dr Meel.
The award comes with USD$40,000 USD in funding, as well as an additional USD$20,000 in Amazon Web Services Promotional Credits to pursue their proposed research plan.
At present, the team has created an initial version of GPUShareSAT, and plan to use the funding to further refine the shared library.
“We are grateful for Amazon's support and recognition of our year-long effort in designing a fundamentally new approach to GPU-based SAT solving. We are excited to explore the frontiers of integration of hardware accelerators and automated reasoning tools,” said Dr Meel.
“We would also like to acknowledge the National Research Foundation, DSO National Laboratories, and the National Supercomputing Centre for their generous support over the past four years,” he added.