Well, welcome to my blog. Most of my posts will be about the project I’m working on: milkyway@home. I’m currently a computer science student at Rensselaer Polytechnic Institute (RPI), doing summer research with RCOS. I’ve been assigned to work on completing the GPU implementation of the n-body simulation that will be running on the milkyway@home platform.
For readers who don’t know and for the sake of completeness, milkyway@home is a distributed computing project run at RPI. Milkyway@home has gathered quite a large following, with over 30000 users with recent credit and over 40000 computers with recent credit (according to the stats page). Enough numbers, let me tell you what I’m working on!
With the original tasks assigned to workers on the milkyway@home platform running out, we need new tasks for the workers to run so they won’t abandon the project. The new tasks will run n-body simulations of the Milky Way and an orbiting dwarf galaxy. The ultimate goal of this project is to better characterize the dark matter distribution in the Milky Way. My task is to implement, test, and release the GPU version of the n-body code. Luckily for me, the GPU implementation is nearly done, unfortunately, the testing is more involved than I expected, not mentioning only having nvidia cards to test on. They also take a while to complete, especially with a larger amount of bodies running on a lower-end GPU.
In the time I’m waiting for the tests to complete, I’m also familiarizing myself with the code. There’s a lot of it. I don’t quite understand how it’s organized yet, but I’ll hopefully get a hang of that soon. I’ve figured out the code that deals with the calculations, as well as some other sections that deal with file handling and such, so I guess the important parts are dealt with.
Thanks for reading, and I’ll hopefully share some interesting material in the future that’s not just related to milkyway@home with you!
