Research Statement

I work on the problem of understanding program performance at a low level, with the practical aim of producing higher performance overall.

Currently I am looking at ways to analyze memory access behavior, with the practical goal of developing tools to help programmers achieve higher memory performance. I'm also interested in the larger problem of analyzing dataflow through interesting architectures, such as GPUs. For these types of analyses, I believe that visualization is an important technique.

In the past I have also studied visualization for Material Point Method (MPM) datasets.

Dissertation

Visualizing Program Memory Behavior Using Memory Reference Traces
A.N.M. Imroz Choudhury
Ph.D. Dissertation, University of Utah 2012
[pdf] [small pdf]

Publications

Topological Analysis and Visualization of Cyclical Behavior in Memory Reference Traces
A.N.M. Imroz Choudhury, Bei Wang, Paul Rosen, Valerio Pascucci
PacificVis 2012
[project page] [paper] [movie] [bibtex]

Abstract Visualization of Runtime Memory Behavior
A.N.M. Imroz Choudhury, Paul Rosen
VisSoft 2011
[project page] [paper] [bibtex] [slides] [movie]

Enhanced Understanding of Particle Simulations Through Deformation-Based Visualization
A.N.M. Imroz Choudhury, Michael D. Steffen, James E. Guilkey, and Steven G. Parker
Computer Modeling in Engineering & Sciences
[project page] [pdf] [bibtex]

Ray Tracing NPR-Style Feature Lines
A.N.M. Imroz Choudhury and Steven G. Parker
NPAR 2009
[project page] [pdf] [bibtex]

Interactive Visualization for Memory Reference Traces
A.N.M. Imroz Choudhury, Kristin C. Potter, and Steven G. Parker
EuroVis 2008
[project page] [pdf] [mov] [bibtex]