Comment by cornholio

Comment by cornholio 15 days ago

You can knock off a zero or two just by time shifting the 700 million distinct users across a day/week and account for the mere minutes of compute time they will actually use in each interaction. So they might no see peaks higher than 10 million active inference session at the same time.

Conversely, you can't do the same thing as a self hosted user, you can't really bank your idle compute for a week and consume it all in a single serving, hence the much more expensive local hardware to reach the peak generation rate you need.

0cf8612b2e1e 15 days ago

During times of high utilization, how do they handle more requests than they have hardware? Is the software granular enough that they can round robin the hardware per token generated? UserA token, then UserB, then UserC, back to UserA? Or is it more likely that everyone goes into a big FIFO processing the entire request before switching to the next user?

I assume the former has massive overhead, but maybe it is worthwhile to keep responsiveness up for everyone.

Reply View 11 replies

cornholio 15 days ago

Inference is essentially a very complex matrix algorithm run repeatedly on itself, each time the input matrix (context window) is shifted and the new generated tokens appended to the end. So, it's easy to multiplex all active sessions over limited hardware, a typical server can hold hundreds of thousands of active contexts in the main system ram, each less than 500KB and ferry them to the GPU nearly instantaneously as required.

Reply View | 2 replies
- apitman 15 days ago
  
  I was under the impression that context takes up a lot more VRAM than this.
  
  Reply View | 1 reply
  
  cornholio 14 days ago
  
  The context after application of the algorithm is just text, something like 256k input tokens, each token representing a group of roughly 2-5 characters, encoded into 18-20 bits.
  The active context during inference, inside the GPUs, explodes each token into a 12288 dimensions vector, so 4 orders of magnitude more VRAM, and is combined with the model weights, Gbytes in size, across multiple parallel attention heads. The final result are just more textual tokens, which you can easily ferry around main system RAM and send to the remote user.
  
  Reply View | 0 replies
parentheses 2 days ago

They probably do lots of tricks like using quantized or distilled models during times of high load. They also have a sizeable number of free users, who will be the first to get rate limited.

Reply View | 0 replies
computomatic 15 days ago

This is great product design at its finest.
First of all, they never “handle more requests than they have hardware.” That’s impossible (at least as I’m reading it).
The vast majority of usage is via their web app (and free accounts, at that). The web app defaults to “auto” selecting a model. The algorithm for that selection is hidden information.
As load peaks, they can divert requests to different levels of hardware and less resource hungry models.
Only a very small minority of requests actually specify the model to use.
There are a hundred similar product design hacks they can use to mitigate load. But this seems like the easiest one to implement.

Reply View | 4 replies
- addaon 14 days ago
  
  > But this seems like the easiest one to implement.
  Even easier: Just fail. In my experience the ChatGPT web page fails to display (request? generate?) a response between 5% and 10% of the time, depending on time of day. Too busy? Just ignore your customers. They’ll probably come back and try again, and if not, well, you’re billing them monthly regardless.
  
  Reply View | 3 replies
  
  nocturnes 14 days ago
  
  Is this a common experience for others? In several years of reasonable ChatGPT use I have only experienced that kind of failure a couple of times.
  
  Reply View | 2 replies
the8472 15 days ago

During peaks they can kick out background jobs like model training or API users doing batch jobs.

Reply View | 0 replies
vikramkr 14 days ago

In addition to stuff like that they also handle it with rate limits, that message that Claude would throw almost all the time when they were like "demand is high so you have automatically switched to concise mode", making batch inference cheaper for API customers to convince them to use that instead of real time replies. The site erroring out during a period of high demand also works, prioritizing business customers during a rollout, the service degrading. It's not like any provider has a track record for effortlessly keeping responsiveness super high. Usually it's more the opposite.

Reply View | 0 replies