Ask HN: Catching Up with Current Datacenters | Hacker News Reader

Ask HN: Catching Up with Current Datacenters

3 pointsDamogran611 comments

I've been out of the datacenter business for...well, not very long. But when I was there it was blade servers and mezzanine networking.

Now I see datacenters with a solid perimeter of LP generators (presumably not for backup) and Nvidia continues to churn out massively more expensive chips and the New mentions 'The biggest AI datacenters ever'

And I'm wondering: what do they look like? What underlying support software/OS/Automation is being used? (Still VMware/Proxmox or something else?)

How doe you get back up to current best practices when theses things have a certain level of 'need to know'?

View on HN

Are you the same guy from here? https://news.ycombinator.com/item?id=44745937

Trying to make another post so your question wouldn't be addressed on the same thread?

If not, would you connect these two threads in a single line? (do the other post relates to your previous feelings or conclusions about the idea?)

I mean, I see that a lot on HN with people doing this with throwaways all the time, but I wanted to be sure.

Not me, but it must be a question of interest, I can check the other thread.

After checking: This is entirely unrelated.

It sounds to me very related. You're trying to jump from "local" to "local datacenter".

If it's not that, can you explain a little better what you're trying to do?

There is no way a single person can amount local servers to compete with Google/Microsoft/etc. If you were led to believe that, you were probably scammed.

This topic (my only one) was a question asking what a current datacenter looks like.

Things like Elon getting called out because there are more generators in his Memphis datacenter than he was permitted for make me thing they're not using the electric grid for power, they're using natural gas?

Nope, you're question was related to "catching up with current datacenters", not "how they look like in relation to energy concerns".

It looks like I'm right (regarding what you're asking, not strictly about being the same person as the other thread), and you're trying to avoid talking about it.

Allow me to paraphrase his ask, which wasn't poorly stated in my opinion. He's not seen an AI DC before, but has some fairly recent general purpose DC experience. What are the differences? How've thing's changed?

Some models often report their training hardware. It doesn't take 5min to figure out that you can't "catch up" with it.

If you have the resources to "catch up", you would probably not ask poorly stated questions on HN, unless you have ulterior motives (which might be harmless, or not).

It seems like a question designed to steer people who might be unaware of such limitations into believing they can somehow buy cheap hardware and make their own datacenter, which is a weird propositon. That's an absurd idea, and such things would only be valuable for cheap spammy purposes (I've been investigating those schemes for a while).

So, we have a few scenarios here:

- Beginner guy who's too lazy to research for himself.

- Scammer trying to dump off hardware that was used for scams into other people's hands (my favorite kind of investigative meal).

- Someone else that is doing the same kind of investigative work I'm doing.

My answer suits these three profiles. It would humble the beginner, scare the shit out of the scammer, and elicit some respect from the investigator.

I know this is hard to follow, but keep focus on why these kinds of profiles are important. If you can think of another one, please share it, and I would rethink my answer in favor of that.

Dude. I don't know how what I asked and what you interpreted got so far apart, and yet I'm compelled to respond.

I've been in IT for more than 30 years. I've worked hands-on hardware for the first 20 of them. There's a datacenter in my office (though it's nearly empty as it's all been migrated to the cloud.)

What does a modern AI hosting datacenter look like? What's the network look like from a compute module to the perimeter of the DC? Is the hardware that hosts AI compute a 6U server with 20 RTX 4090's?

When Elon gets in trouble with having too many turbines...are they generating power at all times because it makes some kind of thermodynamic sense to convert LNG to power, rather than pulling power from the grid?

I'm not talking about the software used in training, I'm asking about the bare metal.

Ah, I see!

If you're not willing to research, you should hire and pay someone (not me) for that information.

Forums are gonna get really quiet if that's what everybody does.

(ChatGPT was massively more helpful than you. Not quite sure if that's great for mankind or not.)

That's not my problem.

Reply on news.ycombinator.com