SUCCEED

   

modelingUniverse 2012
Shodor > SUCCEED > Workshops > Archive > modelingUniverse 2012

To begin the class, Jennifer discussed the different kinds of memory involved in parallel processing. The class talked about Distributed and Shared memory. Jennifer drew diagrams to explain these methods. They also talked about a possible hybrid of these two methods. Then, Jennifer introduced the concept of MPI, Message Passing Interface, that relates to distributed memory and OpenMP, Open-multi processing, which connects to shared memory. She then talked in depth about what the students' code had changed and how the outcome can be altered. Once the students had a chance to see what their code had done, the class moved on to work with specific compiling techniques. Compilers involve writing code that transforms source code one programming language into another computer language called the target language. This meant that the students were going to have to write in multiple ways, but they managed to breeze through it! Jennifer talked a bit more about C, which is the language used for most parallel computing, and OpenMp which are the two new types of code that the students used.

After a short break, Jennifer explained about how to use threads with code in C. This was very interesting for all the students and they had to pay close attention to keep up. Even though it was a lot of material, it prepared everyone to learn about private variables. Finally, the students worked hands-on with MPI and Jennifer explained what all the parts of a piece example code modified along the process.