Amazon Web Services

Filed in EC2 | Game of Life | Linux 1 Comment

As mentioned previously Amazon offers services used by a lot of companies to provide affordable, scalable computing power. That’s great if you’re making money from whatever you’re doing on those servers. I’m not. So, what caught my eye is the AWS free usage tier. This gives you 750 hours of computing time each month for a year. Doing the math shows me that a month has about 750 hours so essentially you can have a server up and running continuously for free. Now this is what Amazon calls a micro instance so it’s not particularly fast or powerful but it’s still useful. I won’t repeat how to set up an account and an instance…you can read about that here. One thing I did differently is to just have a user name and password instead of of a key. No particular reason, just a preference. I can then SSH into the server from any computer or my mobile devices.

Once you’re up and running then you can do whatever you want with it. I uploaded my python modules using scp:

$ scp gol_ver2.0.py chad@54.203.101.174:~/projects/gol_ver2.0.py
$ scp gol_defs.py chad@54.203.101.174:~/projects/gol_defs.py

and then used SSH to log into the machine:

ssh -X chad@54.203.101.174

The -X turns on X forwarding so that I can use the graphical display option. For now though it’s glacially slow even for very small grids that result in small images. I’m not sure what’s going on there but I intend to find out. One thing I’ve tried is using compression and choosing a faster compression algorithm. It turns out that the default algorithm used by SSH is fairly slow. This is simple to change when you SSH in:

ssh -c arcfour,blowfish-cbc -XC chad@54.203.101.174

The arcfour and blowfish ciphers are quite a bit faster than the default though this didn’t really make a difference in the results I got. There’s some other problem slowing it down. I was kind of bummed about this (still am) but it got me to thinking about how to measure whether something interesting is happening in a game of life. That then goes back to the old question of what is life? It’s hard to define but you know it when you see it. Well, even if I could see it I don’t want to spend all my time staring at the monitor. The whole reason for doing this is so I can have a large universe running for long periods of time and just take a look at the interesting things when they occur. I’ll have to figure out how to measure things like the complexity of the grid, self organization and so on.

Autoverse

Filed in Autoverse | Cellular Automata | EC2 Leave a comment

Recently I read Permutation City by Greg Egan and was intrigued by the concept of the Autoverse that plays a central role in the plot. The Autoverse is basically an artificial universe based on a cellular automaton and is complex enough to support a simplified chemistry used to create artificial life. The whole thing runs on distributed computing services that people rent time on.   The book was written in 1994 and in the meantime technology has kept marching on and provided us with an infrastructure very similar to what the Autoverse ran on in the form of the Internet and various computing providers such as Amazon’s EC2 service.

So, I’ll  be doing some projects here to learn more about cellular automata and using Amazon’s computers.

TOP