2018-07-07 17:31:13 -07:00
|
|
|
---
|
|
|
|
layout: post
|
|
|
|
title: "Ray: 0.5 Release"
|
|
|
|
excerpt: "This post announces the release of Ray 0.5."
|
|
|
|
date: 2018-07-06 14:00:00
|
|
|
|
---
|
|
|
|
|
|
|
|
We are pleased to announce the 0.5 release of [Ray][1]. To upgrade to the latest
|
|
|
|
version, run
|
|
|
|
|
|
|
|
```
|
|
|
|
pip install -U ray
|
|
|
|
```
|
|
|
|
|
|
|
|
## Refactoring
|
|
|
|
|
|
|
|
One primary development is a substantial refactoring of the backend, which is
|
|
|
|
currently off by default, but can be turned on by starting Ray with the
|
|
|
|
environment variable `RAY_USE_XRAY=1`. This will be turned on soon after more
|
|
|
|
testing and development and will be on by default in the next release. Some
|
|
|
|
highlights of the refactor include the following:
|
|
|
|
- Merging the "local scheduler" and the "plasma manager" into a single "raylet"
|
|
|
|
process (consisting of a "node manager" and an "object manager").
|
|
|
|
- Faster object transfers by using multiple threads to send objects in chunks.
|
|
|
|
- The introduction of a "lineage cache" to improve latency by removing the
|
|
|
|
global control store from the scheduling critical path.
|
|
|
|
- The removal of the global scheduler (which is replaced by direct node-manager
|
|
|
|
to node-manager communication)
|
|
|
|
- Providing a standardized table-based interface to the Redis-backed global
|
|
|
|
control store.
|
|
|
|
- A preliminary approach for automatically flushing global control store
|
|
|
|
metadata to avoid out-of-memory errors for long-running jobs.
|
2018-07-08 00:38:32 -07:00
|
|
|
|
2018-07-07 17:31:13 -07:00
|
|
|
Note that the refactoring is currently under development and does not yet
|
|
|
|
support fault tolerance.
|
|
|
|
|
|
|
|
## RLlib Improvements
|
|
|
|
|
|
|
|
A number of improvements have been made to [RLlib][2]:
|
|
|
|
- Added DDPG [[1877]](https://github.com/ray-project/ray/pull/1877).
|
|
|
|
- Added a policy server example [[2232]](https://github.com/ray-project/ray/pull/2232).
|
|
|
|
- Multi-agent support [[2286]](https://github.com/ray-project/ray/pull/2286).
|
|
|
|
- Modular policy graphs [[2294]](https://github.com/ray-project/ray/pull/2294).
|
|
|
|
- RNN support [[2299]](https://github.com/ray-project/ray/pull/2299).
|
|
|
|
- Vectorized environments [[2170]](https://github.com/ray-project/ray/pull/2170).
|
|
|
|
- Many bug fixes.
|
|
|
|
|
|
|
|
## Tune
|
|
|
|
|
|
|
|
A number of improvements have been made to [Tune][3]:
|
|
|
|
- HyperOpt integration [[1763]](https://github.com/ray-project/ray/pull/1763).
|
|
|
|
- Many bug fixes and improvements.
|
|
|
|
|
|
|
|
## Pandas on Ray
|
|
|
|
|
|
|
|
Pandas on Ray has moved to the [Modin][5] project and will continue to be
|
|
|
|
developed there.
|
|
|
|
|
|
|
|
## Autoscaler
|
|
|
|
|
|
|
|
The [autoscaler][6] has grown more mature and now supports Google Cloud
|
|
|
|
Platform:
|
|
|
|
- Support for GCP [[2061]](https://github.com/ray-project/ray/pull/2061).
|
|
|
|
- Many bug fixes and additional features.
|
|
|
|
|
|
|
|
## Java
|
|
|
|
|
|
|
|
Ray now supports Java thanks to contributions from [Ant Financial][4]:
|
|
|
|
- Support for Java (e.g., [[2092]](https://github.com/ray-project/ray/pull/2092) [[2094]](https://github.com/ray-project/ray/pull/2094) [[2151]](https://github.com/ray-project/ray/pull/2151) [[2171]](https://github.com/ray-project/ray/pull/2171) [[2225]](https://github.com/ray-project/ray/pull/2225) [[2245]](https://github.com/ray-project/ray/pull/2245)).
|
|
|
|
|
|
|
|
## Other Features
|
|
|
|
|
|
|
|
- A named actors prototype [[2129]](https://github.com/ray-project/ray/pull/2129).
|
|
|
|
- Usage of airspeed velocity for regression testing [[2113]](https://github.com/ray-project/ray/pull/2113).
|
|
|
|
- More efficient serialization of numpy arrays [[1889]](https://github.com/ray-project/ray/pull/1889).
|
|
|
|
|
|
|
|
|
|
|
|
[1]: https://github.com/ray-project/ray
|
|
|
|
[2]: http://ray.readthedocs.io/en/latest/rllib.html
|
|
|
|
[3]: http://ray.readthedocs.io/en/latest/tune.html
|
|
|
|
[4]: https://www.antfin.com/
|
|
|
|
[5]: https://github.com/modin-project/modin
|
|
|
|
[6]: http://ray.readthedocs.io/en/latest/autoscaling.html
|