Raspberry Pi 3 Cluster for FDS

[TH]

Good morning everybody

Anybody has tried to build an FDS cluster composed from Raspberry Pi 3 "computers"? I have seen that Kristopher Overholt installed FDS in one device ( http://www.koverholt.com/2013/01/running-fds-on-a-raspberry-pi/) . Anybody tried to use several of those tiny and affordable linux computers for FDS? 

I have seen some pretty clusters ( http://www.networkworld.com/article/3156748/computers/10-amazing-raspberry-pi-clusters.html#slide1) some of them has 250 Raspberry Pi computers ( 4 threads per raspberry pi). Maybe it could be an affordable solution if performance is good enough ( no comm bottleneck ).

Thanks in advance

Toni

[Ben Ralph]

I have thought about and discussed this a little in the past. I think the prime limitation is memory - with only 1GB per single-board computer spread over four cores it may be lacking. You'd need smaller meshes and more processors and more communication bottlenecks. Having said that - we haven't actually made and tested one.

--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fds-smv+unsubscribe@googlegroups.com.
To post to this group, send email to fds...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/8a803791-a305-4a7f-a54d-bcd3daf77ab8%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Randy McDermott]

FDS can be built on Raspi using the gnu_linux_32 target in the repo.  I thought Glenn even had a set of wiki notes for setting up and running on Raspi, but I can't seem to find them at the moment.

To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/CAK8%2B5_QhR0nxDAahrnFo%2BUQ9PvP1hHZ0xEk-TRhzfG9LvSFcmg%40mail.gmail.com.

[Glenn Forney]

I put some bash scripts in the repo for helping setting up a raspberry.  They are located at bot/Scripts/raspberry.  update_reapberry.sh downloads and installs various packages needed.  setup_torque.sh and setup_queue.sh sets up a queing system.

I tried running parallel jobs using multiple cores on on raspberry and found it wasn't very efficient.  I was amazed that I was able to build and run smokeview.

To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/CAC6wax1BygYVw%3DryCPhZdW%2BMhEZx8j3s_nebPjeLn%2BiNAW1njA%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

--

Glenn Forney