Topics of interest include, but are not limited to:
- theory and application of iterative refinement in different precisions in different applications
- the use of splitting-based approaches (such as the Ozaki schemes) to approximate numerical operations on low precision hardware, including their evaluation and experience in real-world applications
- methods that enable the use of matrix engines or tensor cores for scientific applications
- mixed precision algorithms for different computational motifs (especially those that are not matrix-matrix multiplication or dense linear systems)
- algorithms and software technologies that can accelerate memory bandwidth-limited applications using lower precision formats
- the development of a-priori or a-posteriori error estimators, enabling prediction of the suitability for a workload to leverage reduced precision
- benchmarking the mixed precision performance of HPC systems for scientific computing
- innovative hardware designs that can accelerate or improve the efficiency of scientific simulations.