New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
Vultr has A100, A40 etc CloudGPU hosting.
And there is options like use Runpod(they sells gpu cloud & serverless)
Runpod's Serverless is probably affrodable but depending usecase.
HostKey also offer GPU Cloud too.
Maybe you can look into it.
https://vultr.com
https://runpod.io
https://hostkey.com
https://crunchbits.com/gpu/cloud
https://crunchbits.com/gpu
Not H100, but RTX A6000, 4090, 3090s, etc, available in both bulk and affordable pricing.
Hello
We can do that if you open a ticket support and we will help you.
We had in the past and we still have some customers with GPU with us.
Please open a ticket support with full details
https://alexhost.com/dedicated-servers/
Alexhost
vast.ai
@crunchbits ! Eric's your man!
Vultr: you mostly run a fraction of a gpu, and they are often out of stock. You'll fail at "bulk" requirement unless you contact them directly and arrange something, which defeats the purpose of "cloud". After spinning several machines up and down, they suddenly asked for justification to spin up the next one although I didn't host any web service, blast out emails, ddos, or anything of that nature. Just playing with gpus.
Runpod: you're inside docker container so there are things you can't do. Can't run docker inside docker, can't run anything that requires kernel mod since you don't have access to the host kernel (ex E2E encryption). Any VM/container host has access to the guest instance's entire memory layout. This is true of any provider but with the marketplace model you are ceding this access to god knows who.
yeah, it looks he looking for cheap options.
So I recommend vultr or runpod or hostkey.
If finding best options, he should do business deal with good providers, and get dedicated gpus or get cpus and put to colocation.
It's costs more high but more stable performance with great support.
https://www.coreweave.com/ or https://lambdalabs.com/ might help you