题意：OpenAI Gym 和 Python 线程处理

问题背景：

I am working on a variation of A3C/ACER and I have several workers, each running on its own thread. I am using OpenAI gym environments.

我正在开发 A3C/ACER 的一个变体，并且有多个工作线程，每个线程都在独立运行。我使用的是 OpenAI Gym 环境。

Python threading works fine but it cannot fully utilize all cores. As there are no blocking I/O, it does not context switch.

Python 线程处理工作正常，但无法充分利用所有核心。由于没有阻塞的 I/O，它不会进行上下文切换

I would like workers to somehow to release the GIL while executing actions in their respective environments.

我希望工作线程在各自环境中执行操作时能够释放全局解释器锁（GIL）

I would appreciate your feedback: Does it make sense and it is possible?

我想听听你的反馈：这样做有意义吗？这是可行的吗？

问题解决：

Answering my own question: I found that a quite efficient way is demonstrated in OpenAI universe-starter-agent: GitHub - openai/universe-starter-agent: A starter agent that can solve a number of universe environments.

回答我自己的问题：我发现一个非常有效的方法展示在 OpenAI 的 universe-starter-agent 中：GitHub - openai/universe-starter-agent：一个能够解决多个 Universe 环境的入门代理。

The implementation uses Tensorflow and runs independent processes including a parameter server.

该实现使用 TensorFlow 并运行独立的进程，包括一个参数服务器

I think this can be useful as a reference to other people too.

我认为这对其他人也可以作为一个有用的参考。