Luke Cartey

I am a final year D.Phil student at the University of Oxford. My thesis title is "Domain-Specific Languages for Graphics Cards."

Profile

Over the last five years I have worked in the area of programming languages and programming tools. I have worked extensively with compiler construction tools for developing my own compiler and framework for DSL construction. This includes optimisations for running on the graphics card. I have also worked on projects for programming tools, such as developing IDE plugins in Eclipse for code construction and debugging, as well as code generation and validation from high-level models.

My current research interests lie in how we can generate better code from easier to use languages. I am particularly interested in domain specific languages and generating code for interesting or unique hardware, such as graphics cards. I have worked on and am interested in automatic parallelisation techniques.

Linkedin profile

Publications & Technical Reports

Synthesising Graphics Card Programs from DSLs, Luke Cartey, Rune Lyngsoe, Oege de Moor, PLDI'12, Proceedings of the 33rd ACM SIGPLAN conference on Programming Language Design and Implementation (Acceptance Rate: 19%). [pdf]

Domain Specific Languages for Massively Parallel Processors, Transfer of Status progress report, November 2010. [pdf]

Representing Memory Graphs: Interactive debugging using the Eclipse Framework, 4th year project report, May 2008. [pdf]

Talks

Taming graphics cards for bioinformatics - A talk that was given at BIRC, Aarhus University, 13/09/2012.
Synthesising Graphics Card Programs from DSLs - A talk that was given at PLDI'12, 07/2012.
Automatic Parallelisation: Code Generation for Domain Specic Languages on GPUs - A talk that was given at the 3rd UK GPU Conference in Goodenough College, London, 12/2011.
A language approach to accelerating dynamic programming problems on the GPU - A talk that was given at the 2nd UK GPU Conference in Cambridge, 14/12/2010.
Accelerating Dynamic Programming algorithms using Massively Parallel Hardware - A talk given to the 2010 DPhil Student Conference, November 2010.
HMMingbird: A Hidden Markov Model Compiler - The slides for a talk I gave at the Programming and Design of Hetereogeneous and Reconfigurable Systems workshop in Glasgow, August 2010.