Distributed Processor Load Over Network

[carlosbutler]

I have heard that there is a free app that is able to share your computer processor load over the network to another computer. I have an iMac and would like to see if I could get my MacBook Pro to also work out things, running in parallel.

So if for example I want rip a movie or work out Pi to a billion decimal places (which i dont ) I can have both computers working on this task.

The person that told me has got it set up on their machines, but because this person worked in a shop they apparently weren't able to tell me the program, or what to search for on Google and after my searching around I got nothing.

 

[miles01110]

Generally the program you are using (or more commonly with things like this, the program you write) has to support distributing the load. Most out-of-the-box applications don't, and I've never heard of any program that can tell other proprietary software how to divvy up its required loads over a network. I'd be interested if there was one though, even though its probably a security nightmare.

 

[ScoobyMcDoo]

What you are describing is a cluster. So search for "Mac Cluster" and you will find some stuff out there.

 

[robbieduncan]

Xgrid is built into all Macs since 10.4. As noted above software must be specifically designed and written to use it.

 

[carlosbutler]

well I have been looking around, ended up with http://www.jppf.org/downloads.php , from Wikipedia

Not really sure if this was it, I'll try it when I get back home, although doubtful as to what I was describing.

 

[sammich]

well I have been looking around, ended up with http://www.jppf.org/downloads.php , from Wikipedia

Not really sure if this was it, I'll try it when I get back home, although doubtful as to what I was describing.

It's a framework for Java. Framework meaning it's a set of API's that let you build network processor sharing applications. It doesn't enable any app to magically do things over a network.

Parallel computing is a very difficult programming environment. Some things are easier than others, such as converting movies but that is mostly because you do those in very large chunks, say computer 1 does this half, the other computer does the other half. No overlap or there are huge synchronisation issues. My lecturer told us once: 99% of all parallel code anyone writes will be wrong.

If you're just looking to convert movies, you can just do it manually. It'll save you the headache.