Hacker News Reader

Comment by kllrnohj | Hacker News Reader

> but if it is, I think ARM will take over a large amount of server/cloud market.

Seeing as how Apple has zero presence in the server/cloud market, how do you figure? Are you hoping they bring back xserve?

Otherwise I don't see how Apple having a custom power-efficient chip in a laptop is going to do anything at all in the server/cloud market?

View on HN

If my dev environment and build pipline is all ARM, I might as well get a ARM based server in production if they are available at a reasonable price.

The argument has been that lack of credible Arm desktop / laptops has held back Arm in the cloud. If you can't compile on your desktop then it's much harder to develop / test etc. Now you can buy a Mac and use that to develop for AWS Arm etc.

I don't see your cloud infrastructure throwing everything away and wholesale switching platforms just because you bought a new macbook air, though.

This is all assuming you're doing anything at all that depends on the ISA and pushing prebuilts of it somewhere. Otherwise it's not like deploying Java, Python, JavaScript, etc... changes at all here.

Not throwing things away. X86 will continue to exist for sure. It’s just if ARM is popular to developers, the barrier that sometimes your code don’t run on ARM will eventually be removed. After the barrier is removed, some advantages of ARM will shine, such as more energy efficient, more cores(so when you buy a cloud server you are more likely to have dedicated cores, instead of vcores).

> some advantages of ARM will shine, such as more energy efficient, more cores

ARM doesn't have any such benefits. The ISA has minimal impact on power efficiency, performance, or even core counts.

Apple's specific CPU has so far been pretty good, but that isn't the result of running ARM nor does it translate to other ARM CPUs. The specific CPU implementation matters here, not the ISA.

Sorry but ARM does have an advantage as a 10 year old ISA vs one that still carries the baggage of 40 plus years of development. The exact size of that advantage isn't clear and may be small vs implementation specific features and process differences, but it's still there.

Plus, there have been features of Arm implementations that have clearly given them power efficiency advantage vs x86, big.LITTLE which is now coming to Mac and up until now has not been a feature of x86 implementations.

You've gone from "no impact at all" to "won't throw everything away"

Removing barriers to developers building software on the desktop for the Arm ISA will have an impact on deployment in the cloud. How much - who knows and it's not measurable, but it will.

It's equally plausible, if not more likely, that the impact here is just that Apple laptops are no longer viable for those developing cloud applications where the ISA is relevant.

I've just bought an Intel MacBook Pro so I can continue to develop for x86 in the cloud!

Both things can be true though.

That’s also possible. But at that point I won’t define the new chip as successful.

There are other manufacturers that are building ARM servers these days. AWS even went so far as too build their own chip: https://aws.amazon.com/ec2/graviton/

Reply on news.ycombinator.com