The class started by talking about what a supercomputer is. They discussed top500, a ranking of the top 500 most powerful supercomputers in the world, where they are located, and the performance of suercomputers. After learning about these huge supercomputers, the students got the chance to make their own computer clusters with their tablemates. They rearranged the lab into two groups of 5-6 computers, then cabled the computers together. They then used a program to link their computers together, and in doing so learned some of the basic commands of unix. To illustrate how a cluster can make computing faster, they first ran a model of how the gravitational pulls of stars affects their movement. This model ran rather slowly using only one computer, but when the power of all the computers in the cluster were used it went much faster. The students also realized that when they did computations that required more communication between the computers of the cluster, the model was actually less efficient because all the computers needed to talk to each other.
In the afternoon, the students tested their clusters by running the GalaxSee model. To do this, they ran the model using one computer with a given number of stars. Then they added computers and recorded the time it took the computer to run the model. Then they changed the number of stars and repeated this process. This allowed the students to consider what would be the optimal parameters for running the model on a cluster.
After the break, the students discussed and analyzed their data. This required that they know why they sometimes got outliers, and why multiple tries did not always yield the same results. Then, they looked at a different model and began gathering data for this model using multiple computers. By the end of the day, the students better understood the importance of super-computers, and how multiple computers working together can sometimes complete jobs more quickly.