A Compiler-Centric Query Engine Design for Mixed Workloads and Modern Hardware 11/1/2024, 1:00 PM 2 PM, DBH 3011 Michael Jungmair, Technical University of Munich, Germany Abstract: Relational query engines are increasingly expected to handle more than just relational queries and also run on modern hardware that is increasingly parallel and distributed. However, it is not clear how existing system designs can deal with these two challenges effectively. We propose a holistic, compiler-centric design for data processing systems that is designed for tightly integrated optimization and execution of relational queries, non-relational workloads and user-defined functions on modern hardware. Bio: Michael Jungmair is a third year PhD student at the Technical University of Munich. Supervised by Jana Giceva, he is performing research in the intersection of database engines and compiler technology. So far, this research culminated in the design and implementation of LingoDB (lingo-db.com), a novel query engine based on the MLIR compiler framework

 CloudMapper: A Pay-as-you-go Solution for Accelerating Genomics Sequence Alignment Using Public Clouds Abstract: Single-cell RNA sequencing (scRNA-seq) alignment remains a computational bottleneck in bioinformatics data analysis. As datasets grow in size and complexity, traditional alignment tools, such as CellRanger, face significant limitations, often requiring hours or even days. Furthermore, setting up the necessary infrastructure frequently demands […]