<html><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8"></head><body >Hi Paul and everyone else interested!<div><br></div><div>Thanks for your interest in the raspberry pi cluster project. </div><div><br></div><div>I suggested this project and I've been working out how to do it over the last few weeks. </div><div><br></div><div>I've got a small cluster working and have an SD card image with arch Linux (because it's a minimalist installation without X etc) and the mpich software configured. I've also written some config scripts that can be used to upload rsa keys, programs, etc. </div><div><br></div><div>My original plan was to get people to bring along a raspberry pi / power supply / blank SD card / net cable this Friday so we could have a go at building a bigger cluster.</div><div><br></div><div>However, with the equipment problems we encountered last Saturday at The Ark Nursery, the plan for this Friday is to continue the work at the ark, so we'll have to do the raspberry pi cluster experiment at a later date. </div><div><br></div><div>The extra time will perhaps give me the opportunity to look more deeply into the programming aspects and perhaps I can do a short presentation on how to write a program for the cluster. </div><div><br></div><div>By the way, I've run a Monte Carlo method pi calculation program (using 10^7 points) on the cluster and the results are:</div><div><br></div><div>1 raspberry pi node : ~19 seconds</div><div>2 raspberry pi nodes: ~9 seconds (1/2 time for 1 pi)</div><div>3 raspberry pi nodes: ~6 seconds (1/3 time for 1 pi)</div><div><br></div><div>I guess we can see a pattern emerging...</div><div><br></div><div>Obviously 10^7 points is just so I could test it quickly with the three raspi nodes I have access to. We need to increase the number of points for a larger cluster. </div><div><br></div><div>Hope that helps bring you up to date... </div><div><br></div><div>Malcolm</div><div><br></div></body></html>