Department of Computer Science at UH

University of Houston

Department of Computer Science

In Partial Fulfillment of the Requirements for the Degree of
Master of Science

Amrita Banerjee

Will defend his dissertation

Dragon2011: OpenUH-Based Scalable Visualization Tool for Program Analysis


The size and complexity of real-world application programs can cause challenges when maintaining, parallelizing, upgrading or porting them. In such scenarios, application developers benefit from a structural overview of the code and program details which will help them analyze data usage, control flow and detect code regions profitable for parallelization. Few tools are able to provide such information especially in case of the FORTRAN programs.

Dragon2011 is an interactive system with a powerful Graphical User Interface (GUI) providing a range of information about the structure of source program in a graphical browser-based form, at the level of detail desired. It currently handles programs written in FORTRAN, C/C++ and OpenMP. The features supported include call graph, control flow graph and array region analysis. The necessary advanced program analysis capabilities are provided by the OpenUH research compiler, maintained by researchers of the HPCTools group at the University of Houston. Various phases of the OpenUH compiler extract static as well as dynamic source code information. The latter is enabled by invoking the compiler’s feedback components.

Dragon2011’s predecessor tool, Dragon, had severe drawbacks such as poor layout algorithm, a read-only code browser and cluttering of graphs on the user’s screen. It was developed on MOTIF/LESSTIF with X11 which restricted the tool’s usage to UNIX platforms only. In contrast, Dragon2011 is supported on Windows and Linux platforms. It inherits a scalable graphical layout algorithm from Graphviz. Additional features include – a code editor with syntax highlighting, real-time search functionality and a detachable graph dock. Dragon has been tested on large codes including GenIDLEST CFD application.


Date: Friday, December 2, 2011
Time: 12:00 PM
Place: 550-PGH

Faculty, students, and the general public are invited.
Advisor: Prof. Barbara Chapman