Wednesday, September 1, 2010

My go-anywhere "lab" tools - Scientiae, September 2010

This month's Scientiae topic is research equipment. For me, it's pretty simple. My main equipment consists of my laptop, research group computers, the department computing clusters, and a supercomputer or two. So, yeah, computers. "Lab" is a foreign concept.

This setup lets me work from many locations. In the past few months, I've worked from Wisconsin, Illinois, Tennessee, North Carolina, Virginia, and Maryland (I was too tired in Ohio and refused to open my computer in New York). I can log on to any of the big computers through the internet. Even the data I use is available remotely. I download it to my laptop or directly to one of the other computers from whoever stores it (usually some government organization).

I use the terminal window religiously. I recently added nxclient to my repertoire, which gives me a GUI desktop for an interactive node dedicated to my research group. That's handier than I expected. Between these two interfaces, I can run programs interactively or in batch mode (submit a job that will run when the resources become available).

I reroute the output into a text file so I don't have to babysit the program, I can review the output later, and the program will not crash if my laptop dies or I lose my network connection. If I had to run everything interactively and couldn't record the output, I would have violently and permanently killed at least one computer by now. As it stands, I have non-violently killed several computers several times (including single-handedly jamming the department cluster), but I, my adviser, or the sys admin were always able to bring them back.

And my little macbook is the portal to it all. It's wonderful and dangerous at the same time.

No comments:

Post a Comment