Victorian Partnership for Advanced Computing

Title: Molecular Modelling of Deschampsia antarctica Antifreeze Proteins

Summer Intern: Haritha Kuppam

Supervisor(s): Dave Edwards, Ulrik John, Michael Kuiper, Dev Behera

Discipline: Bioinformatics

Research Objective: Antifreeze Proteins(AFPs) help cold acclimated organisms survive in freezing environments. AFPs from antarctic hairgrass Deschampsia antarctica shares a high degree of homology with AFPs recently isolated and characterized from perennial ryegrass Lolium perenne. And hence the theoretical 3-D beta-roll model of AFP of Lolium perenne
is used as a template model for homology modelling of D.antarctica AFPs. Due to the repetitive nature of the grass AFP sequences, scripting languages such as PERL will be used to search for other possible unidentified AFP sequences within the Deschampsia and other plant genomes.

Final Report: Molecular Modelling of Deschampsia antarctica Antifreeze Proteins

Title: Database of transcriptional regulation in the human genome

Summer Intern: Robert Marshall

Supervisor(s): Robert Flegg, Dev Behera

Discipline: Bioinformatics

Research Objective: The twin goals of this project are to build a
database of a transcription factor binding signatures that allows us to investigate the potential for regulation of every gene and to create a web interface to the data that will allow biological researchers to explore that data.

Final Report: A Database of Transcriptional Regulation in the Human Genome

Title: Drug Resistance in HIV Protease

Summer Intern: Daniel Oehme

Supervisor(s): Chris Birch, Michael Kuiper

Discipline: Lifesciences

Research Objective: This project is aimed at developing models of HIV resistance mechanisms, using a variety of techniques such as molecular mechanics and bioinformatics, which hopefully can be validated experimentally. This project will require detailed analysis of computer modelled drug interactions using a variety of software packages and web based resources.

Final Report: Drug Resistance in HIV Protease

Title: Geometric optimization of turbulent pressure drop in a 2D/3D duct

Summer Intern: Mladenko Kajtaz

Supervisor(s): Chris Seeling

Discipline: Computational Engineering

Research Objective: To develop a tool chain based on ACIS/Hoops/Gambit that allows the automatic optimization of the routing and shape of a duct for the flow of air from a specified inlet to an outlet. The project requires a development of a tool that will exploit geometric modelling with the aim to find a route and a shape of a duct between two points/planes, so a fluid flow through the duct experiences the smallest possible pressure drop. Software packages such as ACIS/HOOPS, Gambit, and Fluent will be used to satisfactorily complete the project.

Final Report: Geometric Optimisation of Turbulent Pressure Drop in a 2D/3D Duct

Title: Develop conceptual design tool chain for branched network turbulent flow

Summer Intern: Stephen Dibb

Supervisor(s): Chris Seeling, Jarrod Sinclair

Discipline: Computational Engineering

Research Objective: The project aims to research the development of a tool chain for the conceptual design of branched networks containing turbulent flow conditions. The tool chain will streamline prototyping of possible network configurations by linking CAD (Unigraphics), meshing (Gambit) and CFD (Fluent) elements. This project was conducted solely as a VPAC CfCP project, however on completion the tool chain would have potential application to several contracted industry projects.

Final Report: Conceptual Design Chain for Branched Network Turbulent Flow

Title: Stochastic Analysis Interface (SAI) Development for LS-Dyna and Fluent

Summer Intern: Ryan Kelly

Supervisor(s): Chris Seeling

Discipline: Computational Engineering

Research Objective: To develop a Python (and/or Java) GUI for the processing of LS-Dyna .k files and Fluent based input files. This project aims to produce a software system for the stochastic analysis of engineering models. Through a Graphical User Interface, the designer will be able to attach to the parameters of a model a probability distribution for their value. They will then be able to run a large number of tests on different instances of the model, the results from which will be collated and presented for analysis. This will give the designer a much better understanding of the behaviour of their model than a single test could achieve. Initially, the project will be developed to support LS-Dyna based model files, with support for Fluent-based files being added if time permits. It is envisaged that support for other formats will be added at a later date.

Final Report: Stochastic Analysis Interface

Title: Conversion of Parallel Directives in the CSIRO Global Climate Model(GCM) to OpenMP?
Form

Summer Intern: Keith Hsuan

Supervisor(s): Robert Bell, Steve Quenette

Discipline: Ocean and Climate Modeling

Research Objective: This project aims to convert the parallel directives in the CSIRO Global Climate Model(GCM) to OpenMP?
directives. This will allow easy porting of the model to (almost) any shared-memory multiprocessors. Also, to identify scope for additional performance gains in the current parallelization, especially related to sections of code that execute serially. The CSIRO GCM code has been previously parallelised using automatic parallelisation directives, however, this has been done using a variety of platform dependent formats. For simplicity, and greater portability, we would like to translate these legacy directives to the OpenMP?
format. After this, optimisations and performance increases will be sought through further parallelisation of the code.

Final Report: Conversion of Parallel Directives in the CSIRO Global Climate Model (GCM) to OpenMP Form

Title: Duplicate checking of large datasets at CSIRO Marine Research

Summer Intern: Roger Ting

Supervisor(s): Robert Bell, Steve Quenette

Discipline: Ocean and Climate Modeling

Research Objective: The project aims to optimize an existing duplicate checking software. The end product should enable future user of this software to handle larger datasets which is not currently possible. It also aims to restructure the existing software which permits easier future extension. The software identifies identical profiles and eliminates duplicate profiles found in the large datasets (as many as 7million temperature/depth profiles) manipulated by the Ocean Observing Networks at CSIRO Marine Research. The existing software is inefficient and difficult to extend.To enable the user to deal with larger dataset, we would like to parallelize existing programs. The new software with streamlined code and better modularity should also enable easier future extension.

Final Report: Duplicate Checking of Large Datasets at CSIRO Marine Research

Title: Interactive inversion software for geoscience

Summer Intern: Tim Robinson

Supervisor(s): Louis Moresi, Patrick Sunter

Discipline: Computational Geoscience

Research Objective: Work with the Nimrod/O development team to develop a prototype interactive inversion front end to the Snark/underworld (VPAC/Monash Cluster Computing) geoscience finite element software package.

Final Report: Interactive Inversion Software for Geoscience

Title: Modelling discontinuity surfaces in planets and moons

Summer Intern: Luke Hodkinson

Supervisor(s): Louis Moresi, Patrick Sunter

Discipline: Computational Geoscience

Research Objective: Working with the Underworld (Monash) and Snark/StGermain (VPAC) teams on simulations of planetary evolution with inbuilt discontinuity surfaces which can be used to represent plate boundary faults, the free upper surface of the planet, or internal boundaries such as the dipping fault surface of a subduction zone. In order to accurately model planetary evolution, certain surfaces must be represented. These include plate boundary faults, the free outer surface of the planet and other internal boundaries. Clearly, not just a geometric representation is required, but also tracking of other physical effects that can occur at these surfaces, such as the transport of fluids and stress. Although already able to perform rudimentary surface tracking, Snark does not possess the infrastucture to handle the effects mentioned above. It is this infrastucture and its integrarion into St. Germain and Snark that is the main focus of this project.

Final Report: Modelling Surface Discontinuities in Planets and Moons

Title: Extending St.Germain

Summer Intern: Martin Elliott

Supervisor(s): Steve Quenette, Patrick Sunter

Discipline: Computational Infrastructrue

Research Objective: StGermain is VPAC's underlying computational framework that forms the backbone to all of VPAC's computational geophysics projects, including Snark, Snac and SPModel. This project aims to extend a collection of useful facilities to a computational coder. The StGermain IO subsystem requires a method of outputting structured data - such as the results of a simulation. This data may be used for visualisation, or further analysis. What is important is that the data is output efficiently in a parallel environment and in a well-defined format independent of machine architecture. To this effect, the HDF5 library will be used to implement the output data files.

Final Report: Not applicable.

Title: Mine Optimisation team with BHP

Summer Intern: Martin Elliott

Supervisor(s): Gary Frowland, Steve Quenette

Discipline: Computational Infrastructure

Research Objective: To assist the mine optimization team at BHP in their development and deployment of linear integer programming models. One of BHP Billiton's business areas is modelling and optimising mining operations to obtain the best value from a mine over its lifespan. The mine models require external input data specific to each mine, and produce a large amount of data about the mining operation. Often, the provided data is incomplete or incorrectly formatted, and so pre-processing is required. Also, the data produced by the optimisation is not in a format best suited for generating reports about the mine operation and so post-processing is required. This project is involved in performing these pre- and post-processing steps to help automate data processing activities.

Final Report: Not Applicable