Educational Advice

CUDA Community,

I’m entering my second year in college and have been following GPU Computing for a couple years now. I’ve spent the past 6 months or so going through some of the resources on the CUDA Training page. GPU computing really has my interest and I can see it being something I pursue.

Does anyone have any long term advice for a college student. I’m not looking for the latest tutorial or best book or anything else i can google.

Any input is much appreciated.

Also, if there is a post similar to this buried somewhere that I’ve missed please point me in that direction.

Implement something. Go to an adviser or teacher and see if you can get some direction for something interesting to run on the GPU. I would recommend taking an image processing class since most of what you learn will be the CPU versions of problems that lend themselves extremely well to parallel implementations. It provides a good basis for getting the mindset. If you take differential equations then check out the PDE solvers that exist on the GPU. Also a linear algebra class is good for looking at matrix problems.

I got into it because I bought a system with a high priced video card and my justification was I would learn to do more than play games with it. I was working on a geostatistics problem at the time and wanted to see how it fared on a GPU. That got me in the door at my current job and that is what I do. Everything I do currently breaks down to image processing, PDE solvers for simulations and ray casting for line of sight models.