Calling C code from Julia | Building a Wrapper | Simple Tutorial

Julia is a great language for speed and efficiency. Often you can implement many basic numerical algorithms with only very little overhead. Still, sometimes you want to call some code you created in another language, for instance c. Either you have access to the source file or just a compiled shared object, Julia allows calling the code with no overhead.

In this video, we look at how to use the Julia functionality called "ccall". I will show how to compile your code as a shared object. In the end, we will create a simple wrapper to make accessing the code even easier.

-------

-------

Timestamps:
00:00 Introduction
00:20 Compiling into a shared object
00:48 Creating a Function Pointer
01:22 Preallocate the arguments
01:43 Using "ccall" on the Function Pointer
02:31 Building a Wrapper
03:41 Outro

[MachineLearningSimulation]

Some Remarks on potential performance improvements:
* When preallocating in the wrapper, we don't have to fill it with zeros. We could do an uninitialized allocation by "output = Array{Float64, 1}(undef, n)"

[CoenArrow]

This was great. Thanks for making it so succinct!

[javadca]

Thanks for nice video. Interesting!
Could you make a video to show how call Julia code from C language?

[orjihvy]

interesting. curious if tensorflow probability will also available in Julia in the future.