In the first project what I did was had the program that run on setonix iterate through a bunch of combinations of parameters and time only the matrix multiplication function for each one. For this project it says to use a maximum of 10 minutes on setonix, while also measuring how big a matrix can be multiplied in 10 minutes. But there is a lot of time used on the setonix node that is unrelated to the matrix multiplication, for example generating the matrices. In addition, wouldn't this mean that each batch run would be testing only one specific combination of parameters. Meaning we would have to run an sbatch command for every single combination of thread number, matrix size, etc.?

Generating the matrices counts towards this time, so you should parallelise the matrix generation. I cannot think of any other time consuming steps. You shouldn't be writing the large matrices to files during your experiments. You are correct, you will need to run `sbatch` for every combination of parameters. The idea is that this should make the sharing of the nodes within the class a bit more democratic -- this way one student cannot use all of our Setonix nodes for several hours, for example. You can still trigger all of your `sbatch` commands from a single script though.