In 1969, humans landed on moon using Apollo-Guidance-Computer (AGC) whose compute power was equivalent to today's scientific calculator. Today we have billions of device which are alteast million times more powerful than AGC lying mostly idle. With this inspiration, have created Ramanujan-Computing which can harness idle computation to run scientific simulations.

There is a project BOINC which tries to solve this problem, however for any new kind of simulation, the developer has to create new binary, distribute that binary to the devices that can provide its computation power. This problem gets solved with Ramanujan-computing. Any device that wants to give its compute-power, just need to install Ramanujan interpreter, and can run any kind of simulation. So, developer can submit any kind of simulation and the idle device can run it without a need for a new binary. On Ramanujan computing have been able to run physics n-body simulation [https://www.youtube.com/watch?v=V5H3jmGoi1k], Phi3 3.8B model inference[https://www.youtube.com/watch?v=EqM3rXKzurE].

This project is MIT-licensed.

Seeking your feedback on the project. And, also, the project is open for contribution. The interpreter is in nascent stage and currently its running more optimally than CPython [15% faster], 20X faster than Octave [matlab]. There is lot to build and help science move forward.