[Trinity College] [Department of Computer Science] [ISG] [Michael Manzke]

Interaction, Simulation and Graphics Lab (ISG)

     
   

Final Year Projects 2002 - 2003

HOME
RESEARCH
PUBLICATIONS
TEACHING
Courses
1BA3
2BA4
3BA5
Current Teaching
2006/2007
Previous Teaching
2002/2003
2003/2004
2004/2005
2005/2006
Final Year Projects
2000/2001
2001/2002
2002/2003
2003/2004
2004/2005
2005/2006
2006/2007
 

 

Supervisor: Michael Manzke

 

All my students are encouraged to use the Concurrent Version System (CVS) and LaTeX for their project and write-up.  

 

Distributed Rendering of Particle Systems


Charles Smith

 

Chromium is a software package for rendering on a cluster of PCs. It facilitates the implementation of parallel rendering algorithms. Chromium is an extension to Stanford's WireGL project. 

Scalable Rendering allows the creation of high performance rendering systems without specialist hardware by utilising clusters of rendering nodes with commercial graphics accelerators. 

Particle systems describe a class of rendering techniques for displaying objects without a rigid structure, including fire, water, and atmospheric effects, which traditional Computer
Graphics approaches find difficult. The term has also been extended to cover decentralised behavioural systems consisting of large numbers of entities which are independent but influenceone another's motion.

Here we describe attempts to extend and improve on some existing computer graphics techniques by making use of parallel computation and rendering, to manage and display complex particle based systems in real-time on modest sized compute clusters.

Charles' Report

 

 

Real-Time Parallel OpenGL Applications on Compute Clusters

 

AnneMarie Walsh

 

This project involves testing an interactive rendering system, Chromium, on a cluster of machines in real-time, using serial and parallel OpenGL applications. It draws together hardware and software systems in an effort to find a viable alternative to expensive custom-built supercomputers, the traditional system required to render intensive graphical applications. Exploiting its aggregate optimised graphics hardware and processing power, Chromium enables parallel applications to execute interactively across a cluster. ROAM, Real-time Optimally Adapting Meshes, is the terrain generation algorithm used for the majority of the testing applications designed. Serial and parallel versions of this program are implemented and their performance with and without Chromium evaluated. The improvement in the frame rate of the applications suggests that Chromium, running on a modest-sized cluster, is a worthwhile consideration for rendering demanding graphical applications in real-time.

AnneMarie's Report

 

 

Open Source Real-Time OS (RTEMS) on SCI Based Compute Clusters

 


 Salman Taherian

 

A new class of real-time application demanding high computation and parallel processing power has begin to emerge. It is believed that previously proven successful compute clusters could once again, offer an optimum and cost-effective solution to such demand. This project investigates the implementation of a specific real-time compute cluster (RTCC). A dedicated cluster interconnect, SCI (Scalable Coherent Interface) - IEEE Approved Standard 1596-1992, with hardware based functionality and deterministic performance was selected as the RTCC interconnect. Implementation
focuses primarily on the hardware based distributed shared memory (DSM) offered by the SCI technology as the basis of inter-communication within the RTCC; while providing a solid platform for support of future SCI functionalities on the RTCC.

RTEMS, Real-Time Executive for Multiprocessor Systems - an "open source", "licence free", real-time dedicated operating system, was evaluated as the target real-time operating system performing on the RTCC. This project investigates the incorporation of SCI technology in RTEMS (supporting SCI functionalities and drivers), as well as implementation of a suitable real-time cluster computing library on RTEMS for support of DSM based RTCC. Finally, this project presents a mechanism of performing real-time computing by utilising the DSM based intercommunication. Synchronisation and mutual exclusion facilities are also supported through implementation of a suitable two stage lock mechanism.

Salman's Report
 

 

 

Design and Evaluation of a FPGA based Microprocessor Project Board

 


Ross Brennan

 

This project should evaluate a possible upgrade of the CPUs used for the 2BA4 microprocessor project. After all the current 68k CPU represents early 80's technology. This is not necessary a disadvantage for a 2nd year class, but the component will be soon obsolete. The project will assess a FPGA solution with a customisable CPU for a number of reasons.

The CPU could not only be used for the 2nd year microprocessor project but also for subsequent courses. This would allow student not only to simulate their design but to implement and test it in hardware e.g. processor unit (alu + shifter + fast registers) design and simulation, instruction processor design, memory management units, translation look aside buffers (TLBs)... It would give a great degree of flexibility.

The project will most likely evaluate a LEON core on Xilinx XCV2V1000-5FG256C FPGA and a Virtex-II PrototypeBoard.The HW-AFX-FG256-200 board supports 256 pin FG packages and is therefore suitable for the XCV2V1000-5FG256C FPGA. For more information see: "FPGA CPU Links".

The European Space Agency developed the LEON core, a SPARC* compatible integer unit for future space missions. It has been implemented as a highly configurable, synthesisable VHDL model. To promote the SPARC standard and enable development of system-on-a-chip (SOC) devices using SPARC cores, the European Space Agency is making the full source code freely available under the GNU LGPL license. The model comes with a generic testbench and test program, and includes support files for the Synopsys VSS and Modelsim simulators. It also features a built-in disassembler for debug purposes. The VHDL model is fully synthesisable and contains synthesis scripts for Exemplar Leonardo 1999.x, Synopsys FPGA-Compiler, Synopsys-DC and Synplify.

Ross' Report

 

 

Parallel-Octave

 


Jill Louise Brangan

 

 

This project will modify Octave source code to allow for the parallel execution of numerical computations performed by the Octave interpreter. GNU Octave is a high-level language, primarily intended for numerical computations. It provides a convenient command line interface for solving linear and nonlinear problems numerically, and for performing other numerical experiments using a language that is mostly compatible with Matlab. GNU Octave is also freely redistributable software. You may redistribute it and/or modify it under the terms of the GNU General Public License (GPL) as published by the Free Software Foundation.

 

 

There is a link to the departmental final year project page here.

 

Final Year Project Coordination

[Trinity College] [Department of Computer Science] [ISG] [Michael Manzke] Last modified:  23rd September 2006

Free counter and web stats