Cloud computing environment that supports Julia

General Usage
1

Problem

I am trying to see the how much GPU could help to accelerate my code by ArrayFire.jl. However, I currently do not have access to GPU on my machine. I have tried two different ways

  • Try to use the integrated Intel GPU. However, the following error keeps bumping up and I am not sure how to resolve this issue
    image
  • Use some cloud computing environment. But the commonly used Google Colab has not supported Julia yet, I am not sure if there is some good alternative I could use since trying them one by one could be extremely time-consuming.

Could someone help me, thank you in advance.

2

Haven’t tried it myself but maybe this helps? Julia on Google Colab: Free GPU-Accelerated Shareable Notebooks

3

If the problem is worth a few cents to you and you have a credit card i can recommend. vast.ai
They are an PC with GPU peer-lending platform which has lower rates than other cloud services since they offer consumer GPUs.
To run julia on them select the image
[nvidia/cuda:9.0-cudnn7-devel-ubuntu16.04 ](https://hub.docker.com/r/nvidia/cuda/)
And download the Julia binaries for Linux. I’ve had some JuliaGPU tests running there and they worked fine. You then have to install all the packages.
If you add a credit card (the only way to pay) you get $2 of free budget which is a lot when instances cost between 0.06$ for a 1x GTX 1070 and 1.60$ for 8x RTX 2080 Ti per hour, you can also bid for a cheaper price but then you can be interrupted at any time.
They also offer access to GPU machines from Google and Amazon at their prices.
Important things:
Do not forget to turn of or destroy it over the web interface.
If it is turned of you are only paying for storage.
On some instances you also have to pay for bandwidth.

Disclaimer: I got a $10 credit for reporting bugs in there interface to them. I think they are cool guys.

4

If using Nvidia GPUs, you might find it easier to use CuArrays.jl, which is integrated more into the Julia compiler.

5

I had never heard of them before. It is good idea but I wonder how NVIDIA feels about its consumer GPUs are used in such a setting. Maybe this is the reason the website does not seem to provide any information about people working there? Or perhaps I was not able to find it?

6

@hckr
NVIDIA has little to feel about. Peoples private homes (where these PCs supposedly stand) are not data centers.
However the creators were quiet active on reddit. There also is a Discord Server.