Erlang-OTP and the Parallella Board
parallella.org
parallella.org
The 64 core version of the board never shipped in quantity and has apparently been abandoned. There still isn't a practical way to connect them beyond the networking provided by the ARM processor. Lastly, there isn't a lower cost revision of the board intended to use in clusters, which is a shame because it's that ARM/FPGA chip that represents the bulk of the cost & heat of the Parallella board.
I love the idea and for the most part the community seemed like reasonable folks... it's just that the Adapteva folks were never really able to deliver on the promises made in the original kickstarter and so the whole ecosystem pretty much stagnated.
This morning I read an article about a new many core project using the 64 bit ARMv8 which is supposedly going to be a topic for a talk at this year's Hot Chips...
http://www.kitguru.net/components/cpu/anton-shilov/phytium-t...
It's not that surprising that the community is moving slowly: They've shipped something not that far above 10,000 boards, many of which are likely in the hands of Kickstarters backers that bought them because we liked the idea more than because we have time to do all that much with them.
They need another iteration and another magnitude increase in community size, really.
For example, I bought an Oculus DK2. I've pretty much not done anything with it because I didn't realize at the time how much computing power I needed to really have fun with the DK2. When I received it and tried it out with my MacBook Pro, I was like "this is really cool and I'd love to play with this more, but I'm not shelling out at least a couple of hundred if not almost a thousand dollars for a desktop with a reasonably powerful GPU that would make the Oculus fun to play with.
LeapMotion was another device where I knew several people that had pre-ordered one and ended up doing nothing with it. One of my colleagues has had a Myo sitting on her desk practically unused.
Companies making hardware for developers need to find a way to figure out who is actively working with their devices and contact those that are not to let them know that there is a way to sell their unused devices to someone else who might want to develop for it.
Though you have mechanisms like port drivers and NIFs to interact with the outside world, they don't seem to have been very popular for doing systems programming or using low-level POSIX interfaces, even if there is some work on that front. It'd be interesting to exploit Erlang's unique properties in building lower level OS infrastructure. I should probably play around with some of the bindings that have been written one day.
Also worth checking out is the Nerves Project, a buildroot-based embedded Erlang distribution: http://nerves-project.org/
I'm not entirely sure what you're trying to ask.
Or, more succinctly, don't get this thinking you're getting a 16 core Erlang node, not exactly.
RPi2 has a Quad-core CPU and costs $35 - easy to put together a cluster. Runs standard Erlang - simple to experiment with concurrent programming. Built a web-service that delivers 100's of pages per second on a single device - low memory usage - just great.