Anyone have an idea for a group project?

Well something i want to try is getting multiple raspberry pi's together and then get the thing where you can have multiple raspberry's running as one to try and combine the power and run a windows machine or a high end linux distro computer.

Well something i want to try is getting multiple raspberry pi's together and then get the thing where you can have multiple raspberry's running as one to try and combine the power and run a windows machine or a high end linux distro computer.

Kyle,

  It's a nice idea, and several people have tried it, but it doesn't work very well.

There's no good way to "combine the power" to run "as one".  In particular,

you can't combine the memories into one large memory, and you can't speed

up a sequential job such as emulating a windows machine.

If your job can be split into a number of small jobs that don't need much interaction,

(known as embarrassingly parallel) then you might see a speed up compared to

a single RPi, but the network speed is only 100 Mb, so that limits the amount of

communication you can do between RPi's.

You should do the math, comparing the price/performance of a RPi to an

x86 PC for some benchmark you are interested in.  You will probably find

one PC is less expensive than the equivalent number of RPi's (including

all the necessary power supplies, cables, SD cards, network hub, etc.)

See for example:

I hear that hadoop is incredibly slow on pi from various blog posts, and yes pls lower your hope as the speed is really appalling.

http://raspberrypicloud.wordpress.com/2013/05/03/running-hadoop-java-and-c-word-count-example-on-raspberry-pi/

Well something i want to try is getting multiple raspberry pi's together and then get the thing where you can have multiple raspberry's running as one to try and combine the power and run a windows machine or a high end linux distro computer.

Kyle,

  It's a nice idea, and several people have tried it, but it doesn't work very well.

There's no good way to "combine the power" to run "as one".  In particular,

you can't combine the memories into one large memory, and you can't speed

up a sequential job such as emulating a windows machine.

If your job can be split into a number of small jobs that don't need much interaction,

(known as embarrassingly parallel) then you might see a speed up compared to

a single RPi, but the network speed is only 100 Mb, so that limits the amount of

communication you can do between RPi's.

You should do the math, comparing the price/performance of a RPi to an

x86 PC for some benchmark you are interested in.  You will probably find

one PC is less expensive than the equivalent number of RPi's (including

all the necessary power supplies, cables, SD cards, network hub, etc.)

See for example:

I hear that hadoop is incredibly slow on pi from various blog posts, and yes pls lower your hope as the speed is really appalling.

http://raspberrypicloud.wordpress.com/2013/05/03/running-hadoop-java-and-c-word-count-example-on-raspberry-pi/