prime

make get-deps - download dependencies

make compile - compile

make run - run in foreground

make daemon - run in background

make stop - stop

make progress - get current progress in %

make save - save results in text file

Input N-value and output filename are specified in prime.config

TODO

To make it work on several nodes (servers) the straightforward solution is:

• On startup we must take some time for other nodes to connect
• Connection from other (slave) nodes can be implemented similiarly with make target and escript, sending net_adm:ping
• Most code from prime.erl will go to a manager_server.erl which will manage each respective node
• Master should start supervised master_server on each slave node with help of erlang:nodes command
• Computing can be started after some timeout or manually
• Gathering of results can be implemented with gen_server:multi_call to managers

In order to make production system for similar task we should minimize usage of Erlang's internal message passing capabilities and use fault-tolerant message passing network service like RabbitMQ or other AMQP protocol implementation.