I.e. a cloud instance that behaves exactly like a normal GPU instance, where behind the scenes the GPU is shared. The advantage here is you only pay when the process is actively using the GPU (not the whole time the instance is running). The downside is the instance would take ~25% longer to run any GPU task.
Ask HN: Would you use a shared GPU cloud tier?
1–10 of 20 posts
Re: Ask HN: Would you use a shared GPU cloud tier?
#2Maybe prefix with "Ask HN:"?
Re: Ask HN: Would you use a shared GPU cloud tier?
#3Yes, I would use it if the price is affordable.
Re: Ask HN: Would you use a shared GPU cloud tier?
#4Maybe prefix with "Ask HN:"?
Fixed, thank you!
Re: Ask HN: Would you use a shared GPU cloud tier?
#5Depends on your workload...
Re: Ask HN: Would you use a shared GPU cloud tier?
#6Isn’t this just the cloud? You pay for what you use
Re: Ask HN: Would you use a shared GPU cloud tier?
#7Isn’t this just the cloud? You pay for what you use
Essentially yes. The difference would be that we measure "what you use" more granularly since you'll only be charged while something is being run on the GPU, not the entire time the instance is running.
This is where the "shared" aspect would come in, since during unused GPU cycles someone else could use the GPU.
Re: Ask HN: Would you use a shared GPU cloud tier?
#8Depending on the price difference from a standard GPU instance, absolutely.
Re: Ask HN: Would you use a shared GPU cloud tier?
#9Yes, I have wanted something like this for a while. I try to avoid using gpus where possible because of the expense, and the ephemeral nature of my use.
Re: Ask HN: Would you use a shared GPU cloud tier?
#10Sounds like an extremely complex technical problem. I also suggest to look at the use cases when this is needed. One of the problems is that loading weights into the GPU will be so slow that it will be really hard to share the GPU between different processes - causing long time to offload and load. Would love to learn more about what you do.