Introduction to High Performance Computing - Jacobi coursework

This code implements the iterative Jacobi method to solve a system of linear equations. See the Wikipedia page for a full description of the Jacobi method.

Compiling and running

The code can be compiled by typing make. To change the compiler or flags, you should modify the Makefile.

The program can be run without any arguments to solve a default problem. The -n and -i arguments can be used to control the matrix size and maximum number of iterations. For example, to solve for a 500x500 matrix, use the following command:

./jacobi -n 500

Use --help to see a full description for all of the command-line arguments.

On BlueCrystal, you should obtain your timings by submitting a job to the queue:

qsub jacobi.job

Modify the jacobi.job file to change the arguments passed to the program.

Sample runtimes

Here are the runtimes that we achieve with the starting code for a few different matrix sizes.

Matrix size	Solver runtime	Iterations	Solution error
500	1.48 seconds	1465	0.02495
1000	10.9 seconds	2957	0.05005
2000	130 seconds	5479	0.09998
4000	1180 seconds	10040	0.1999