Pollux is a Message Passing Cloud orchestrator designed to distribute and manage complex algorithms, including NP-Hard and Complete problems, across multiple Cloud-based machines.
In today's world many computational problems from network routing to semiconductors to AI are becoming more and more complex.
We believe that the next step for optimization of such complex grand scale computation is to leverage the massive parallelization the Cloud can offer.
We believe that this step has not been reached yet, as there is no genuine Cloud suitable framework for this type of computation; most of the frameworks used for similar purposes are actually meant for supercomputers, not the Cloud.
Our goal is to provide such a framework, designed for Cloud architecture while easy enough to use for engineers from any field. Through a clear interface they should be able to parallelize any algorithm without being Cloud experts.
Our goal is not to treat the Cloud as a super computer, it is not. Our goal is to virtualize a super computer on the Cloud and finally open these capabilities to a wider audience.
Pollux Message Passing API is currently limited to C++ but can be extended to Python, Rust, ...) in the future.
Hi again, I've updated the README with a minimal set of informations. I hope this clarifies the purpose of the project. I'll later work on the detailed API documentation.
All feedbacks are welcomed.
Thanks for the feedback sacheendra.
You are right about the lack of documentation, starting with a proper README, definitely on the top of the TODO list.
Still working on testing the structural Verilog language support, then will soon focus on documentation.
If there are developers around willing to early test structural verilog support or Naja: C++ netlist data structure, don't hesitate to contact me.
One thing to note: Naja-Verilog and Naja interact but are not tied: Naja structural verilog parser can be used to build any data structure or in any project needing structural verilog support (Yosys or Vivado synthesis outputs for instance).