SUCCEED

   

parallelUniverse 2016
Shodor > SUCCEED > Workshops > Archive > parallelUniverse 2016

Today the students made a bccd cluster using their computers and ethernet cables. They used this cluster to run some programs that would otherwise take a long time to run on their individual computers. After hooking up the ethernet cables, they proceeded to connect all of the computers together. They took a look at how many cores were at their disposal. They then talked about what the cores are used for and they connected their cores and set up the network. The penguins on the screen showed them how many cores were at their disposal. Then they were all assigned an ip address from the head node. They then learned about nodes and created their passwords and learned some commands. They looked at some programs in bccd. Then they ran galaxsee on their computers. Some commands were run to show them some information about galaxsee. The students compared how fast the program ran on different amounts of cores, and then discussed how the program runs on the computer. They ran an experiment to test how the computer runs galaxsee and how the processes were spread out throughout the cores. From this experiment, they learned that the more cores there are, the slower the program runs. After break the students ran a program called pandemic, which can be run on just one core or can be run on multiple cores. They looked at and changed the different parameters of the model. After, the program was run on multiple cores, the students were able to run the simulation with many more cores. They learned that for certain sized simulations, you could run it on different amounts of cores. Then they ran Sieve, which is a model of how to find prime numbers using an algorithm that looks at a number. It also looks at this number's multiples. This program can also be run on both singular and multiple cores.