• 设为首页
  • 点击收藏
  • 手机版
    手机扫一扫访问
    迪恩网络手机版
  • 关注官方公众号
    微信扫一扫关注
    公众号

juliamatlab/julia-matlab-shuttle: Matlab interface for calling Julia

原作者: [db:作者] 来自: 网络 收藏 邀请

开源软件名称(OpenSource Name):

juliamatlab/julia-matlab-shuttle

开源软件地址(OpenSource Url):

https://github.com/juliamatlab/julia-matlab-shuttle

开源编程语言(OpenSource Language):

MATLAB 84.6%

开源软件介绍(OpenSource Introduction):

julia-matlab

This is a Matlab interface for calling Julia, an open-source language providing many features of a Matlab-like environment but offering many of the performance benefits of C. The primary purpose of this Matlab-Julia connector is present a way to enhance performance in Matlab without the need for writing MEX files.

NEWS

Since the author doesn't use this himself, this repository has not been maintained. If you're interested in getting it working, you'll need to start from a version of ZMQ that existed at the time this repository worked: https://github.com/timholy/ZMQancient.jl

It's also very likely that changes in julia will require additional updates.

If you're willing to put the work into it, you are free to take "ownership" (this is released under the MIT license).

Installation and configuration

First, you need ZeroMQ installed on your machine. ZeroMQ is an efficient cross-platform library (available for Windows, Mac, and Linux) used for the communicaton between Matlab and Julia. It is assumed that this is installed as a system-wide library. On (K)Ubuntu this can be installed simply using "apt-get install libzmq1". At present, version 3 of ZeroMQ seems to be too buggy to use, so you should make sure you're installing from the stable (2.x) branch.

Second, from within Matlab navigate to the "mex" directory of this repository and execute the make_mex script. This will compile the necessary MEX files. You need a compiler on your system for this to work.

Finally, configure your communications with Julia by running the julia_configure script. This will allow you to define shortcuts for both local Julia instances and instances running on remote servers. Naturally, you also have to have Julia installed, on every machine that you plan to use.

Using julia-matlab

For a local connection (which is the recommended way to start), you can launch Julia from within Matlab using the juliastart command. After that, read the help for juliacall, juliaparse, and juliasetvar. These are the three main commands that let you send data to the Julia session, perform operations on it, and return the results. Be aware that Julia will hold on to variables between calls, so you can set up fairly elaborate computations with repeated calls.

For a remote connection, you'll need to launch Julia on the remote machine, and then run the zmq_server_julia function. If you've changed any of the default ports in the Matlab julia_configure script, make sure you set the right values. For example, if you prefer to use port 5556, launch the server this way:

julia> load("zmq_server_julia.jl")

julia> run_server("tcp://*:5556")

Errors

Errors are typically reported back to the Matlab client. However, they are also mirrored on the command line of the Julia server. For that reason, if you have trouble you may prefer to launch Julia by hand rather than letting Matlab launch it for you.

If communications between Matlab and Julia get interrupted, an easy fix is often to kill the Julia instance and restart.

Limitations

It would be nice to allow CTRL-C to gracefully interrupt and recover the Julia communication. However, at present this is not implemented.

The most important part of the communication is the serializer (in juliaserialize.m). This targets Julia's native serializer. However, be aware that this is probably not the best solution, and there have been proposals to use a more standards-based serializer such as Thrift or the IPython Notebook. If you are thinking of using this repository as a model for targeting Julia from another language, you are advised to consider first working with the Julia community to implement a more standards-based serializer.




鲜花

握手

雷人

路过

鸡蛋
该文章已有0人参与评论

请发表评论

全部评论

专题导读
上一篇:
cwxcode/LSTM-matlab: LSTM by matlab发布时间:2022-08-17
下一篇:
lawrennd/mocap: Matlab software for processing motion capture files.发布时间:2022-08-17
热门推荐
阅读排行榜

扫描微信二维码

查看手机版网站

随时了解更新最新资讯

139-2527-9053

在线客服(服务时间 9:00~18:00)

在线QQ客服
地址:深圳市南山区西丽大学城创智工业园
电邮:jeky_zhao#qq.com
移动电话:139-2527-9053

Powered by 互联科技 X3.4© 2001-2213 极客世界.|Sitemap