How to: build a graphics cluster

Question

Currently, I am waiting for my second RPi to arrive and then I will begin my fun building a RPi cluster. Also, later I want to extend it more and more. =)

I am building it at home, home-experiment, just for learning, nothing more. =) But I also thought about some use-cases of this cluster also but I am not a system administrator or an engineer and I have no clue whether this can be done in some "efficient" ways or not.

So here is what I would like to achieve:

I have a laptop with Win7 which has a somewhat cool videocard inside but I thought it might be a cool experiment if I can move the graphics processing from my laptop or from specific apps to this cluster.

For example: When I am using Illustrator, Cinema 4D, Autodesk Maya, etc. onthe laptop I want to be able to lend the graphics processing to the cluster on the same network instead of the GPU or CPU.

I am assuming that this requires a lot of work on the laptop too. Of corse it might not make any sense to people but it seems fun and has a good experimental value. =)

So is this possible to achive? Is that possible to lend this cluster the heavy processes from another devices on the same network?

Regards, Ben

Answer 1

Using your pi cluster to parallelize a closed-source PC application seems very difficult. First of all, some piece of the application must run on the Pi. For open source applications which run on the Pi and already have cluster support, it is quite possible.

For Blender (as @InkBlend says), there is multiblend, which runs rendering in parallel on a cluster. Also, since version 2.6, Blender natively supports network rendering. I would, however, not expect great speed-ups, a program using several cores on the laptop will perform a lot faster than the PIs, especially if they utilize the GPU for computing too. For Blender, GPU rendering is in development. The GPU on the Pi is quite powerful, but is hard to use for general computation since OpenCL is not (yet?) supported.

Another possibility is distributed compiling using distcc.

I think it might be more fun and educational to try to write some parallel program on the Pi cluster, for example using MPI, than trying to use the PIs to speed up a PC application. See the Pi supercomputer @InkBlend mentions.