search

PyOCL / OpenCLGA

A Python Library for Genetic Algorithm on OpenCL
MIT License
119 stars 34 forks source link

distributed GA #15

Closed john-hu closed 7 years ago

john-hu commented 7 years ago

As a user, I would like to connect all of my devices to run a single GA model.

kilikkuo commented 7 years ago

As now I could setup pyopencl env in my laptop & VM. I'd like to design the framework. Basically, I'll leave most wrapper & packer empty, and focus on the Inter-Machine-Communication stuff, first.

Host

  1. Problem divider (divide the problem into smaller ones with different initial argument values)
  2. Wrap this oclGA program & sub-problem into an item.
  3. Dispatch these items to multiple machines
  4. Results collector ( Wait until all items are done )

Client

  1. A light-weight listener to receive the sub-problem item.
  2. Executor ( Unwrap the item and execute )
  3. Results packer ( To pack the results into problem-specific format )
  4. A light-weight notifier to notify the result package.
kilikkuo commented 7 years ago

I think this issue could be closed by 1) c8ceab07570baa67c22d040a9eacb7267c2a79fc 2) dedaf675b0e3e12fe85e3f67347523ee23262232 3) https://github.com/kilikkuo/py_simple_host_target/