Hacker News Reader

Comment by p2t2p | Hacker News Reader

In the end of the day it will be Apple disrupting x86 on servers. I already can run all my work related stuff natively on M1. There’s JDK, Node is here as well, ones lots of devs have M1 on their desktops it will only make sense to run aarch64 on servers as well.

View on HN

I do admire how Apple did the M1 release. There was a lot of publicity with developers, good performance and solid roll-out with availability in everything, and it seems to be picking up steam. It seems obvious to everyone they are fully committed and that if you spend the time getting your software ready for it, then you will have a market. I'm not even interested in the Apple ecosystem and I am familiar with the major developments.

> In the end of the day it will be Apple disrupting x86 on servers.

Apple made it happen on the client, but Amazon is making it happen on the server. The people who want to run their own DCs or hardware will end up part of the niche left on x86. I don't see Apple doing anything in the server space beyond providing minis like they do now.

I think what's interesting is that having access to a good development platform (the M1 mac) might enable projects like Docker, Node or Linux to get up to speed on ARM earlier. The dev experience is really much better on a local machine than working through SSH on a server.

From my standpoint there isn't much else to speed up. I'm already running a bunch of workloads on AWS Graviton2. Having an m1 Mac did help a bit with testing, even though the final is running on ARM Linux. Which btw, Linux already runs on ARM just fine from what I've seen.

I’m no expert in the area, there must be someone in Apple doing a profit analysis on selling server cpu even if the profit is tiny the economy of scale alone carries a benign

https://www.idc.com/getdoc.jsp?containerId=prUS47529021 says the server market is about 12 million servers in 2021. I don’t think you can call that “economy of scale”, compared to 200 million iPhones.

(I’m not sure whether that includes servers at large cloud providers, but I don’t think that matters. Unless there’s a huge performance advantage, I don’t see Amazon, Google or Microsoft buying Apple server CPUs)

I’d expect Apple already has more than one proof-of-concept M1-based operational data centers.

If they are to be commercialized, I’d expect them to be part of a development stack related to Xcode, or some kind of AWS like services for applications developed for the Apple ecosystem.

Offering generalized compute would invite too much initial cost performance comparison and they couldn’t keep up with demand.

Whatever goes in the Mac Pro will likely be as capable as x86 data center chips at a fraction of the power requirement. Whether they put it in a server is another question

iServer with 64 cores and 128 GiB DDR4 for only $ 399,999.95

iServer with 64 cores and 256 GiB DDR4 for only $ 699,999.95

They have the Mac Pro rack: 28 cores, 384 GiB DDR4 for only $19,699

You are off by an order of magnitude :)

Adjusting for inflation, that's quite a bit cheaper than an Apple Lisa was.

Do you think a lazy troll is contributing to the conversation? You could look at Apple’s old XServe pricing to see how much markup they historically had, or compare the current MacPro to equivalent workstations — it won’t be 2 orders of magnitude, and based on what we’ve seen so far it’s more like to be within the range of +-25%.

Remember Xscale, anyone?

I got to use an Intel XScale in a product once. I never did see the point of it's choice.

Xserve

Sorry you are right. Xscale was an Intel cpu :)

And one with an ARM instruction set at that!

Reply on news.ycombinator.com