Totally feel you on this. I’ve got 1500 users at our org using Apple systems, and once the ARM products are out and we have to start replacing aging systems with those, they’re going to be seriously screwed. We’re still not 100% compatible for many products (I’m looking at your lame ass, Adobe) and this will only exacerbate the issues. If it were up to me, I’d simply ban them. Saves us a ton of headaches in terms of support and infrastructure. Alas… I have not such power.
And that is why I got out of the tech repair field and now manage software licensing. Because yeah… keeping up just wasn’t fun any more after twenty years of doing the work (and thirty plus years of being a PC wonk).
In older times actually they cared a bit. I rememer that some M68k Macs had an optional secondary processor board with a 65c02 to emulate an Apple //c, had also a modified video card to better emulate the older PC.
Another interesting story is that an emulated the Macintosh on the Amiga 3000 was faster than the original.
Then Apple decided to leave the Personal Computer arena to go in the more lucrative smartphone market.
Because if you buy a 1200 EUR PC you’re a bloody nerd especially a desktop, but if you but the newest EUR 1200 smartphone you’re hip.
And the transition from 68K to PowerPC was actually kind of a train wreck. They wrote a 68K emulator for PowerPC and created the Mixed Mode Manager which handled transitions between 68K and PowerPC code. Most of MacOS was running in emulation for years.
This is a really interesting development, and one that should not be surprising if you’ve been paying attention to hardware. The current ipad pro already beats any intel laptop chip in single-threaded performance, and utterly smokes it in performance-per-watt as well.
But that’s not the whole story - the writing is on the wall at the high-end, too - Amazon just released their second-gen ARM-based servers for AWS, and in fact, I’m working on moving Boing Boing over to them right now - they are cheaper (because they are vastly more power efficient) and have better performance handling usual LAMP stack tasks - in fact the BB workload is about 25% better on the ARM servers running centos 8 aarch64 than on intel.
The other interesting part of this is going to be overall graphics performance. The Intel integrated graphics are laughable compared to what the Apple chips have already. This is going to put every laptop apple sells base graphics waaaaay above a comperable windows laptop, and both can also choose to use discrete chips still at the high end.
I game on my macs, so I’m especially eager to see how they benchmark. The LAMP stack already works on aarch64, and so does xcode, which means homebrew will as well, so that’s all the major pieces moved over for development, and since Apple showed of the Adobe Suite and office running natively already, they’re pretty far along.
What I think will be more interesting now is what Microsoft and Google do next. Chromebooks on ARM are obvious, but I bet Microsoft would love to move to dual-arch too, so their tablet and PCs can unify some.
Yes, at this moment, Boing Boing runs on AMD hardware, As having more cpu cores with 10% lower performance but 25% better power-per-watt is a better choice for the workloads Boing Boing has. Interestingly ARM flips that, as the power consumption goes down another 15% but performance is higher for the workloads we have.
I think the bigger story here is what the implications of Intel dropping ARM development is going to mean for them in the longterm. The writing has been on the wall for years now - even back in 2013 when I was at Wikimedia, they were considering then-aarch32 chips to replace intel, and that was quite a long time before Apple showed just how far you could push that arch.
Forced workflow changes suck, but you may end up in a better place because of it. I switched up my workflow a few years back to use a home VM server for the grunt work. Desktop hardware is powerful enough to easily run a handful of Win/Linux instances for compile and testing.
It took some adjustment time(and forced me to give up some bad habits), but the end result is far more flexibility than I used to have. I’m largely client-agnostic now; I can code on anything with a screen if I have to
Switch to Linux now, and you will be fine, old man.
Parallels versions have been releases roughly on a yearly basis, and you pay a yearly fee for Pro and Business licenses anyway. So Parallels and its users won’t give a fuck.
Classic was to run pre-OS X apps in OS X. It was literally a virtualized instance of OS 9 that seamlessly projected those applications into the OS X UI. What you’re thinking of was called the far less catchy “Mac 68k Emulator”.
Rosetta 2 purports to make the switch seamless but things like plugins won’t work since you apparently can’t mix native with emulated code in the same process. This can be a big problem with DAWs and other plugin-centric applications.
Apple definitely has a good track record here so it’ll be interesting to see what happens.
Long-term, yes, but I’m not so sure in the short-to-medium term. They announced continued support for intel for at least three more years, but more importantly, I’m not sure they have a replacement for Xeon-class compute for things like the Mac Pro / iMac Pro (and if they do, I really want to see it!) If that’s true, we may see intel support continue for some time just to support pro customers.
It’s funny how all you guys haven’t updated your facts or opinions from like 20 years ago. For example, the newly announced iOS 14 will run on iPhones dating back to the 6s, released in 2015. For comparison, Google updates their phones to new OS releases for 3 years (including security updates!). As another point of comparison, I’m running the very latest Mac OS release on a 2012 Mac Mini. It will not get the newly announced macOS 11 upgrade, but that’s not shocking for an 8 year old computer. It IS still getting security updates, I just installed one yesterday.
Yeah, that makes me nervous. I have a 2019 iMac and use Steam for my occasional gaming needs, mostly native on MacOS but sometimes with Parallels. I’ve invested hundreds (thousands!?) of Euros into my steam library. I don’t play demanding stuff but strategy games like Total Warhammer 2 or Xcom 2. As far as I understand it I can kiss it goodbye or buy an additional windows gaming pc (will not happen). I have no interest in playing iOS games on my Mac.
I hope for a last Intel iMac coming out so I can upgrade and wait the situation out.
You seem to be the one here spouting predictions with no evidence. They already showed demos of some pretty compute heavy applications working fine. They have a track record of doing this well when they switched to PPC and when they switched to Intel. One big difference this time is they are actually building custom silicon, which might make doing this translation/emulation even more effective than before. So sure, there’s always an outside chance that they’ll screw this up, but I feel pretty justified in trusting an organization with a long history of doing this well more than I trust a rando on Boingboing.
From the keynote, it seemed like x86 applications would get converted to ARM on installation. So it would be an un-optimized ARM application, not an emulated program. So installation may take a while before you can use the application, but I would think it should be faster launching and on feature use.
Keynote queued to the claim about translation at install time:
I think you have good reason to be optimistic. One huge difference between now and when Apple was on PPC is scale. Back then, Intel shipped orders of magnitude more x86 chips than PowerPC chips, so they were in a position to invest an order of magnitude more in R&D, which made it very difficult for PPC to keep up.
These days, Apple ships an order of magnitude more ARM-based chips than Intel does desktop chips, because of how many iOS devices they are already shipping. So they can easily match/exceed Intel’s R&D investment. Switching the Mac over just increases that advantage.
For the same reason, I think you should be MORE optimistic about gaming than before. Now that a single codebase can easily target both Macs and iOS devices, I think it will be easier for game developers to justify spending significant amounts of time developing and optimizing game engines and games.
Not really. I only experienced the 68k to PPC transition first hand and the performance was abysmal. I worked in a research lab with a 68040 mac next to a PPC mac, and despite the PPC having a dramatically faster CPU it was unbearable to use with almost any 68k app. I have heard the intel transition was better, and I expect the ARM one to be better yet – software design today is much more suitable for emulation as it is naturally more removed from the underlying hardware and binary translation and emulation has advanced a lot since the 90s. So I expect this transition will be between “good enough” and “hardly noticable” at least as long as it is supported. However, Apple isn’t going to support this forever. They have a pretty good history here so you can probably expect quite a few years of support but eventually you won’t be able to run old apps on a modern version of MacOS without third party tools like a VM.
One thing is they could now be free from whatever contract they have with Intel and then pick the x86 platform that performs better for the price. An AMD ThreadRipper or EPYC Mac Pro would have been interesting and more price competitive.
Honestly the CPU isn’t really that important. Developing a game that works well on both a tablet and desktop computer is far more difficult that developing an engine that runs on both ARM and Intel.
