from Hacker News

Ask HN: Can you get started on GPU programming without having one?

by karolsputo on 10/25/22, 12:19 PM with 4 comments

I’m looking to get into GPU programming (especially Nvidia) to expand my AI knowledge and get deeper into the actual training and AI frameworks.

My problem: I don’t own a GPU.

I have a M1 MacBook so connecting a Nvidia GPU is currently out.

Does anybody know how one might get into stuff like that on the cloud?

I’m getting a feeling that I’ll run into so many problems and restrictions that I’ll have to wait until I have enough to build a PC (doing that anyway, just wanted to get started sooner..)

  • by cheptsov on 10/26/22, 7:16 PM

    There is a lot of options:

    Large clouds (rather expensive): AWS, GCP, Azure - consider using spot instances that are cheaper

    Free GPU platforms: Colab, Kaggle

    ML platforms (cheap GPU): Jarvis Labs, Lambda Labs, TensorDoc, Genesis Cloud

    In case you decide to go with AWS, check out dstack.ai, an open-source CLI to run anything on AWS GPU

  • by waynecochran on 10/25/22, 12:35 PM

    I do a lot of GPU development using AWS / EC2 / CUDA. If you don’t need tensor cores you can find cheap instances. Helps to have good shell programming skills and a good text editor like emacs. You can use visual studio with an ssh connection if you prefer.

  • by bjourne on 10/25/22, 7:28 PM

    Google Colab is one option.

  • by giantg2 on 10/25/22, 3:22 PM

    Maybe a Jetson Nano?