Julia IDE, GPU and GUI packages

juno
gpu
vscode
#1

Hello,
I’m back after long break to Julia language and need some advice about IDE.
During the time of my break I programmed some web development(html, css, Sass, Javascript, Ajax, JSON, Fetch API, React, DOCKER … ) and I used vscode as my IDE and it was really good thing. In Julia before the break I used Jupyter Notebook to work with Interact, but it wasn’t satisfying for me. Now I’m planning to make GUI for some complicated calculating program in which I want to introduce some “drag and drop” interface however in my calculation which are really time-consuming i plan to use GPU programming methods(parallel programming) and it’s really important for me to have high performance. If it goes about GPU programming I’ve been reading “CUDA by Example: An Introduction to General-Purpose GPU Programming” by Jason Sanders and Edward Kandrot, but I don’t want to use it because I don’t want to limit to only NVIDIA GPU and also read some about OpenCL.
So my questions:
1)Which IDE?(I thinked about JunoLab or vscode extension however i heard there was some problems with Julia 1.1.0 integration) for Julia.
2) Which package for GUI, to do these effects with good performance, because I want my application to do fast calculations and show results(it must be useful, I don’t want too much visual effects).
3) Which packages for GPU programming would you recommend?

#2

The VS Code extension problems with Julia 1.1 are fixed, so that is no longer a problem.

6 Likes
#3

What about GUI/interface packages. Except Interact I also used Makie package, but there were problems with dependencies packages, however I managed to add this package. But if there was any better package that anyone can recommend I would be grateful. What about Gtk or Genie?
And what about GPU programming?

#4

It can be usefull, I have HD Graphics 5500 in my actual notebook however I have second notebook with Radeon(don’t remember which version but it’s probably better). Is that mean that I can’t use CUDA-based packages?

#5

Another GUI option is immediate mode GUI which is fast but maybe more GPU resource consuming.

https://github.com/ocornut/imgui

1 Like
#6

yes, CUDA is a vendor-only SDK.

1 Like
#7

That’s very interesting. And it have Julia package here: https://github.com/Gnimuc/CImGui.jl, that only difference is the syntax(which for me looks more friendly than the original syntax). I’ll check it for sure.

#8

The last thing to obtain is GPU package and hardware choose. I have to notebooks, one with Intel HD Graphics 5500(GPU) and Intel® Core™ i3-5005U CPU, the second with AMD Radeon HD 8280(GPU) and AMD E2-3800 1,3. In both I have 4 GB RAM.
Ofcourse I aware of that I use very weak stuff, so for the nearest time I think about some new graphic card up to 50 US dollars.