Friday, December 05, 2008

Hello Ubuntu

The new computer arrived, but in the meantime I've been given a to-be-scrapped computer to use as a testbed for my "linux project". What I'm hoping to do is build a linux cluster. At first it'll just be two computers, but I hope to expand it as things go to surplus.

Why? Because with the amount of information we'll be acquiring (especially once we start 454 sequencing in earnest), I'll need the computing power for processing the data. I'd like a 64 bit computer, but from a Windows standpoint that means using Vista, and the government doesn't currently allow Vista on their computers without seeking exceptions ... and that paperwork is a bit of a pain. Besides, several of my programs don't work on Vista yet, which means I'd need to keep an XP machine around anyways (rendering the above moot). *sigh*

Anyways, if I cluster a couple of linux machines, I can achieve (I believe) similar results. So, anyone build a cluster recently? Also, what bioinformatic tools do you use regularly on linux? I've already installed Geneious and the Bio-Linux base programs (which is mostly Emboss) and Artemis.


Anonymous said...

I haven't built a cluster... but I would have thought that if you want to run things in parallel, how fast each node can talk to each other is very important (depending on how you parallelize your programs) and you won't get talk hapenning at supercomputer like speeds. That said, why don't you apply for supercomputer time instead?

kcsphil said...

Have you tried Greg Laden's blog at

He writes a lot about linux, and if he can't help you I bet he can point you to folks who can.

Thomas Joseph said...

Anonymous: Thanks. I guess the issue boils down to accessibility and cost. Geneious allows for access to a computer cluster (3000 processors) in New Zealand using their "Green Button" feature. It doesn't come free though.

Phil: I'll get in touch with him. Thanks.