Swarthmore College Department of Computer Science

Andrew Danner: I/O-efficient Algorithms

andy research image

My current research interests are in I/O-efficient algorithms (also called out-of-core or external-memory algorithms). On data sets larger than the amount of available internal memory of a computer, the transfer of data between slow hard disks and faster internal memory, not CPU speed, limits computing performance. Working in a theoretical model that mimics this behavior, I am interested in finding efficient ways to solve problems in computational geometry on large data sets. I also look at possible applications in geographic information systems (GIS). The STREAM project page summarizes some recent contributions by my research group to hi-resolution elevation data analysis and modeling. While my emphasis is on theory, I prefer to develop solutions that are practical enough to implement and be applied. One example of a practical project that I work on is TPIE—a templated, portable I/O environment written in C++ that makes it easier for people to develop I/O-efficient applications. I also contribute to the open source GIS community through the GRASS project.