Hacker News new | past | comments | ask | show | jobs | submit login

How could a cloud GPU that's a specific model be any different in speed from the same thing in your bedroom, if you could get nvidia to even sell you one in the first place?



The consumer GPUs are much faster, cheaper and more available than most of the cloud GPUs.


Be specific about which models you're comparing.

One of the big deals about data center GPUs is that they have a lot of RAM, more than any consumer card.


Availability on the market vs your wallet. Power consumption vs how much power you can safely draw from an outlet in your apartment/home/place before a breaker trips. Massive waste heat vs your native climate and whatever your place is capable of offering (i.e. AC).

Anecdotally, having a single high powered modern GPU in my office generates enough heat that it is in part why I keep a temperature monitor on my desk to ensure I'm not cooking myself. Having more than one gets far more interesting.


They're probably alluding to the large number of GPU models available to retail customers which are not allowed in datacenters.


Then they're completely missing the point of the cloud GPUs.

If you're not running something that needs NCCL et al, yeah, go run something under your desk.


4090s support NCCL and are not allowed in datacenters


Cloud GPUs aren't better.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: