ADVANCED SCIENTIFIC COMPUTING RESEARCH WITH WAFER-SCALE ACCELERATION THIS PROJECT AIMS TO ADVANCE ENERGY-EFFICIENT, HIGH-PERFORMANCE COMPUTING (HPC) BY OPTIMIZING CEREBRAS’ WAFER-SCALE ENGINEER (WSE) TECHNOLOGY TO ADDRESS DOE AND NETL PRIORITIES, INCLUDING CLIMATE MODELING, CARBON MANAGEMENT, AND ENERGY CHALLENGES. THE OBJECTIVES INCLUDE DEVELOPING AND BENCHMARKING KRYLOV SUBSPACE SOLVERS, SUCH AS CONJUGATE GRADIENT AND INCOMPLETE LOWER-UPPER (ILU) PRECONDITIONERS, FOR SINGLE AND MULTI-WAFER ARCHITECTURES, IMPROVING SCALABILITY, ENERGY EFFICIENCY, AND TIME-TO-SOLUTION FOR HIGH-PERFORMANCE COMPUTING (HPC) APPLICATIONS.