Message Boards Message Boards

0
|
25192 Views
|
5 Replies
|
3 Total Likes
View groups...
Share
Share this post:

Raspberry Pi cluster useful for Mathematica parallel computing?

Posted 9 years ago

Hello,

I own a Raspberry Pi B+ and all the stuff you need for it and I also installed Mathematica. Today, I heard from the option to sum up the calculation power of several computers in Mathematica. Furthermore, I watched a video about creating a cluster with the RPi. So, is it possible to have more calculation power in Mathematica by creating a cluster of RPis? Has anyone experience with this (if it's possible)? Does anyone know a good tutorial?

Thanks, Kutsubato

POSTED BY: Richard S
5 Replies

Check this out https://youtu.be/i_r3z1jYHAc It is a cluster system using Raspberry Pi

POSTED BY: Jose Calderon

I bought a Raspberry Pi 2 specifically for this purpose. Note that its not in what commands are available for efficient parallel processes, but how you divide the workload I just got my first on which i plan to optimize some code i need to perform faster, and as a test for what this thing is capable of, everntually i will scale to 2 then 4 Pis, and who knows from there

POSTED BY: Steven Yampolsky

I'd love to hear about your results, please let us know!

POSTED BY: Sam Carrettie

The version of Mathematica bundled with the Pi does seem to support network parallel computing, so this should be possible (Go to the Evaluation > Parallel Kernel Configuration menu.) However, note that only a handful of functions are natively integrated with Mathematica's parallel processing abilities. I'd love to hear if you make any progress with this, though!

POSTED BY: Jesse Friedman

Parallelize is general and can be ran on custom defined code. So "natively integrated" functions such as ParallelMap is just syntax convenience (if you meant that). For example, search for Mersenne primes:

Parallelize[Select[Range[9000, 10000], PrimeQ[2^# - 1] &], Method -> "FinestGrained"]

{9689, 9941}

Also it is often useful just to run many independent computations in parallel on a cluster.

POSTED BY: Sam Carrettie
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard

Group Abstract Group Abstract