Scalable compute services.
Simple object storage.
Run managed Kubernetes clusters.
Tools & Integrations
Automate your infrastructure.
Deploy pre-built applications.
Connect, share and learn
DevOps and development guides
Questions & Answers
Development and systems Q&A
by Justin Ellingwood
Powered with NVIDIA GPUs would be awesome for machine learning
It would be great but with the latest Eula from nvidia it’s not legal to use GeForce class videocards in data center for tasks like ML and AI; only quadro and Tesla are allowed.
I would rent to use GPUs powered database anyway
The new optimized droplets might be what you want.
Super cool news, Keith! Thanks DO!
Same! Am using DO for most of my cloud server work. Adding GPU adds the machine learning feature well for DO! :)
Adding consumer oriented GPUs could be a really big win. I would look at Nnvidia's 1070/1080/1080TIs/Volta. Amazon is not in a great situation with their offer of AI oriented hardware at the moment (actually they have never been). They offer only enterprise stuff which are not always the most efficient way to satisfy this niche. For instance I get 3-5 times better results with a single 1080TI vs AWS p2.xlarge ( which costs ~750$/month, same as the 1080TI).
BTW, I really like the effort DO is putting to compete with AWS. Would love to see where all this goes.
Great suggestion! I'm a Product Manger here at DigitalOcean working on Machine Learning and I. I just wanted to send a quite note to let you know that we are researching GPUs right now. Also, be sure to check out our Machine Learning One-Click by looking into https://do.co/ml.
More to come soon,
Any news on this topic? Amazon is offering way to expensive stuff.
Any info would be much appreciated! Knowing that the NVIDIA policy towards the data-center is quite restricted, so probably will not see GPUs like 1080/2080 in DO https://www.theregister.co.uk/2018/01/03/nvidia_server_gpus/
I have used AWS gpu instances. It's too expensive and not suitable for small applications and tasks. So it would be awesome if Digital Ocean could offer GPU instances..
So any update on this ?
Any update would be greatly appreciated. Desperately in need of a droplet to which I can add CUDA GPUs as now I need to go to AWS to work with them...
AWS GPU instances are extremely costly. I am waiting for DO to launch their GPU instances for my applications and tests.
You won't be notified about changes to this idea.