Laura Toma

I am a Professor of Computer Science.
Department of Computer Science
Bowdoin College
8650 College Station
Brunswick, ME 04011
Office: Searles 219

Publications | Student research | Teaching | CV | About me

My research is in the theory and practice of cache-efficient algorithms for large data, and in particular applications that involve large, high-resolution data in Geographic Information Systems (GIS). Together with great students we explore algorithms for fundamental problems on terrains such as visibility, flooding, sea level rise and least-cost-path surfaces. Our goal is to come up with approaches that are resource-efficient (CPU, IO, cache, parallel), are backed by algorithms that we can theoretically prove efficient, and at the same time work well in practice. Ultimately, our goal is to transfer these algorithms into free and open-source software. I am grateful for the past support of an NSF award 0728780 (2007-2013) which enabled me to develop this research with students at Bowdoin. Memory-efficient algorithms and parallel algorithms share many techniques and insights which brought me towards exploring high-performance computing using Bowdoin's HPC grid. My DBLP page | Google Scholar page.

I finished my Ph.D in 2003 at Duke University, Department of Computer Science. My thesis advisor was Lars Arge. My dissertation focused on IO-efficient algorithms for modeling flow on very large terrains (terraflow | terrastream), as well algorithms for basic graph problems like IO-efficient breadth-first search and depth-first search, IO-efficient topological sort, IO-efficient minimum spanning trees and IO-efficient shortest paths.


In Spring'23 I am teaching Computational geometry (csci3250) and Algorithms (csci 2200).

Other classes I taught (see teaching for links to course websites): Introduction to Computer Science (1101), Data Structures (2101), Algorithms (2200), Computational geometry (3250), Algorithms for GIS, Spatial data structures, Computing with massive data.

(Some) Projects

See publications for a complete list.