The organisation of the reading material for MN320 is as follows:

BACKGROUND WORK

1. Make sure you are familar with UNIX and Matlab using the appropriate source material.

2. Read the introduction to html - the first part of the MN320 assignment is to get up a small blurb about yourself on the web.

3. Then read all the material on the web on the history and overview of scientific computing in section 1.

4. Next read the background material on computer architecture and networks in section 2.

WWW NOTES

5. We will go through the web notes of Chapter 3. You will be required to do the assignments from Sections 3.1-3.4 and 3.8.

6. MATLAB WORK. The next component of the work is to solve laplace's equation (as described in the assignment) using Matlab. You may find some of the material in section 2 on Partial Differential Equations useful.

7. FORTRAN WORK. The second component of the assignment will involve getting the matlab code running on the University supercomputer using Fortran90. This will necessitate reading all of the Guide to SGI Power Challenge in section 6 as well as material in section 5 on Fortran90.

8. PVM WORK. The third component of the assignment will be some programming in a distributed computing environment using PVM. Section 4 should be read.

9. VISUALISATION. The final component of the work will be visualisation of outputs. More web material will be set up at a later date on the use of Khoros.

SUBJECT MATERIAL We will be working through the Parallel Linear Algebra Book. The first 3 chapters are background material. Chapters 4 and 5 are multiplication of matrices and data lay out in a parallel machine. Chapters 6,7 and 8 are on Gaussian elimination for dense, banded and sparse matrices. Chapter 9 is on parallel iterative techniques. In the Appendix there is some useful material on BLAS and Lapack which will be needed for the assignment.