Apple Filed for 'Apple Rosetta' Trademark in Japan Earlier This Year

[JMO1]

No. Apple cannibalizing themselves is an obvious smart move. You just need to get users locked into one of Apple's products and they'll end up being part of the ecosystem. Students will start wanting the rest of Apple's product lineups since they work well together.

i actually agree, not a pipe dream.If their growth is come if from services, they now want more people in the platform. To do this is the lower priced options. Think iPhone SE at $399

 

[cmaier]

How much trouble is in recompiling really? Compiler backends have already taken care of it, and LLVM is especially great at it.

It’s a lot of trouble if the developer won’t do it. And some apps will never make the leap. Some apps use Sdks that may not be supported. Some developers no longer exist or can’t be bothered. etc.

And most importantly, who is going to buy a machine not knowing which software you currently use runs on it? Rosetta would be A fantastic safety net.
[automerge]1592837800[/automerge]

PPC->Intel make sense since Intel CPU's were much powerful than IBM/Motorola making "rosetta" working like a breeze.
I don't see this coming with ARM. Many people are busing Mac's for Bootcamp only and others (like me) are using Parallels to virtualize Windows which can use only limited resources to gain 1:1 performance anyway. Making the ARM decision Apple may loose lot of users who value virtualization/Windows compatibility. I still hope those CPU's will not be fully ARM but perhaps x64/ARM hybrid which would make more sense anyway

Apple’s chips are far more powerful than Intel’s, so makes sense this time too

 

[Tekguy0]

Everyone wondering about performance, but it's probably all about price. Imagine an education focused MacBook for $499. Cheaper than Chromebooks if the MacBook lasts for more than 2 years.

Most schools use Chromebooks until support runs out or they are so physically worn out that they cease to reliably function (trackpads, keyboards, and SSDs failing). After their lives as chromebooks, they get Linux and serve to teach kids coding. Most will hit about 7-8 years of life before being sent to China so some poor kids can remove precious metals from them. Schools don't have a ton of $$$ to spend anyways, private schools notwithstanding, so keeping CBs as long as possible is standard in most Public/Charter schools. Also Apple needs to make them manageable and more durable, as with chromebooks each students account can be individually managed, and MacOS does not have such as user-management and asset-management system. Finally, Retina Screens look nice, but is very expensive to fix. A chromebook screen costs ~$25, and a case assembly with KB and Trackpad costs about the same, and many are spillproof as well. The only real competitor is the base model iPad.

 

[ddtmm]

Everyone wondering about performance, but it's probably all about price. Imagine an education focused MacBook for $499. Cheaper than Chromebooks if the MacBook lasts for more than 2 years.

It's just the processor that's changing, not the entire hardware platform. I can't see how they can cut the price that much - let alone why they would in the first place. If Apple wanted to release Chromebook-like notebooks they could have used i3s, Pentiums or cheap AMD chips. I wouldn't hold my breath on Apple releasing anything inexpensive. Ever.

 

[Detnator]

Oh no. Here we go again. I remember that this was how you had to run Power PC apps on Intel machines. That was living hell. What a kluge. If Apple is going to make us do the same for new ARM computers, I am going to be done with it and get a PC.

What on earth are you talking about? It was as seamless and transparent as imaginably possible. How was it a kludge and living hell?

 

[justperry]

A computers computer network And the computer software in which development of the computer programmes on a global communication network and the download for translating and performing are possible, computer software for computings performed by a cross platform, computer software, Electronic machines [apparatus and their parts]

Who writes this cra.....

 

[cmaier]

Who writes this cra.....

Presumably a bad auto-translation?

 

[iStorm]

Oh no. Here we go again. I remember that this was how you had to run Power PC apps on Intel machines. That was living hell. What a kluge. If Apple is going to make us do the same for new ARM computers, I am going to be done with it and get a PC.

True, but that was 10-15 years ago, which is ancient in the technology world. Virtualization and emulation has improved greatly since then, so hopefully it wouldn't be quite as kludgy.

 

[cmaier]

True, but that was 10-15 years ago, which is ancient in the technology world. Virtualization and emulation has improved greatly since then, so hopefully it wouldn't be quite as kludgy.

It really was pretty good even back then. I returned to Macs in the leopard time frame and about half the software I used most frequently was PowerPC code. I seldom had any problem running any of it, the speed was good enough that I usually didn’t notice, and really the only giveaway was that the UI didn’t quite match (which had nothing to do with the emulation).

 

[DaRev]

Return it if you’re too worried, but PPC chips were supported for at least 5 years after the Intel announcement.

If you are using this professionally I would stick with the Intel version knowing all the software really works, and it might be a year or 2 before all your apps are fully supported, especially if you run anything from the smaller devs.

If this is just your Mac and you got it because you wanted the best thing available, return it and get one of the new ARM ones if they are even announced at WWDC or a later date.

 

[mrxak]

To this day, I still know Mac users keeping 10.6.8 machines around because they were the last ones to have Rosetta support. I have one myself. If Apple does switch to ARM, maybe that will be okay in the long run, if they can get all the other computer companies switching to ARM (I've always been a RISC > CISC guy). But in the long run, we'll see Apple drop Intel support entirely in a very short number of years, and a ton of software will get left behind. It will just fragment the ecosystem even more.

 

[Jyby]

Science Rules!

 

[Gasu E.]

Obviously the key question here is going to be performance. How fast will binary translation/emulation be and just as importantly how fast are Apple's ARM chips going to be relative to modern Intel Core/AMD Ryzen.

The original Rosetta's performance was acceptable (quite impressive actually on a technical level), particularly for Apple's notebook lineup, but this was due in large part to just how far behind the PowerPC chips used in Apple's notebooks were compared to x86 at the time. Even the original Core Duos used in the first generation MacBooks absolutely destroyed not only the PowerPC G4s used in Apple's iBook/Powerbooks, but even a lot of PowerPC G5 based Macs.

I highly doubt Apple is going to have quite that large of a performance margin this time around so it will be interesting to see just what they can achieve.

...OTOH, if you put an SSD in an old (circa 2010) Mac with a pre Sandy Bridge Core Ix chip (or even Core 2 Duo) it is surprisingly usable today, so perhaps, for casual users this won't matter as much...

Can't say I'm excited. Please let this just be an additional ISA to be used where it makes sense, and not a whole scale transition (unlikely I know, but I can dream right.)

Caveat: My Computer Science expertise is 30+ years out of date. But it seems to me the performance issues you see when using a CISC architecture to emulate a RISC architecture, could be much much less when you use a RISC architecture to emulate a CISC architecture. I wonder if someone with more current knowledge could address this.

 

[lostinlodos]

...I highly doubt Apple is going to have quite that large of a performance margin this time around so it will be interesting to see just what they can achieve.

...

It’s funny to read all the Rosetta comments here and elsewhere. Some people thought it was amazing. Others say it was a cluster.
porting/recompiling shouldn’t be much of a problem for anyone in the Apple ecosystem; that is licensed developers using apple’s tools for Apple distribution, etc. And excluding external independents.
thebreal question is did Apple somehow take the ARM system and bring it forward 2 or three generations to Be on equal footing with current desktop and server chips. Be it other RISC designs or AMD64.

 

[justperry]

Presumably a bad auto-translation?

I'm not convinced, I've seen similar language in American patents.

 

[cmaier]

I'm not convinced, I've seen similar language in American patents.

I read a LOT of patents. They aren't this bad unless they are PCT applications from a foreign application.
[automerge]1592845004[/automerge]

Caveat: My Computer Science expertise is 30+ years out of date. But it seems to me the performance issues you see when using a CISC architecture to emulate a RISC architecture, could be much much less when you use a RISC architecture to emulate a CISC architecture. I wonder if someone with more current knowledge could address this.

Nope, it's bad either way. But not insurmountable.

 

[funkboy]

Obviously the key question here is going to be performance. How fast will binary translation/emulation be and just as importantly how fast are Apple's ARM chips going to be relative to modern Intel Core/AMD Ryzen.

I highly doubt Apple is going to have quite that large of a performance margin this time around so it will be interesting to see just what they can achieve.

IMHO it'll depend on whether Apple decides to go the easier route of "upgrading" their existing A-series iPad SoCs, or if ARM Macs get a clean-sheet design that's more in the Neoverse class of cores, along with additional silicon that improves execution of emulated amd64 instructions.

 

[farewelwilliams]

the first gen ipad didn't use an apple designed chip

Wrong. First gen iPad used the A4 chip. It literally did use an Apple designed chip.
[automerge]1592852397[/automerge]

Most schools use Chromebooks until support runs out or they are so physically worn out that they cease to reliably function (trackpads, keyboards, and SSDs failing). After their lives as chromebooks, they get Linux and serve to teach kids coding. Most will hit about 7-8 years of life before being sent to China so some poor kids can remove precious metals from them. Schools don't have a ton of $$$ to spend anyways, private schools notwithstanding, so keeping CBs as long as possible is standard in most Public/Charter schools. Also Apple needs to make them manageable and more durable, as with chromebooks each students account can be individually managed, and MacOS does not have such as user-management and asset-management system. Finally, Retina Screens look nice, but is very expensive to fix. A chromebook screen costs ~$25, and a case assembly with KB and Trackpad costs about the same, and many are spillproof as well. The only real competitor is the base model iPad.

No. Chromebook replacement cycles are 4-5 years. Don't know where you're getting your data from. My aunt is active in the LAUSD school system and they change out their Chromebooks every 4 years. Seems like you're just guessing there.
[automerge]1592852465[/automerge]

It's just the processor that's changing, not the entire hardware platform. I can't see how they can cut the price that much - let alone why they would in the first place. If Apple wanted to release Chromebook-like notebooks they could have used i3s, Pentiums or cheap AMD chips. I wouldn't hold my breath on Apple releasing anything inexpensive. Ever.

There's an overhead for even using lower end Intel chips, whereas Apple could simply reuse an existing chip and exiting fabrication lines with 0 royalties.

 

[justperry]

I read a LOT of patents. They aren't this bad unless they are PCT applications from a foreign application.
[automerge]1592845004[/automerge]

Nope, it's bad either way. But not insurmountable.

Ok, you win..Haha, no seriously, I didn't even think about it being a translations, thanks for pointing that out.

 

[Spectrum]

Any further news on this following today’s announcement? Would the software Apple described enable an older macOS to run?

Unless someone releases a highly compatible (and performant) x64 emulator for Arm, there's none. So far, this has not happened, and who knows what the future brings. But I would not expect such a solution in the short or even mid term.

 

[RandomDSdevel]

Any further news on this following today’s announcement? Would the software Apple described enable an older macOS to run?

It was in the keynote, mentioned by name. They're calling this iteration 'Rosetta 2,' with that new version number, though. Some of the details under the hood are different. I don't know if any articles cover it yet.

 

[Spectrum]

It was in the keynote, mentioned by name. They're calling this iteration 'Rosetta 2,' with that new version number, though. Some of the details under the hood are different. I don't know if any articles cover it yet.

Yeah, what I mean is, would Rosetta 2 enable an older macOS to run via a VM? Or would it only work to Translate software?

 

[Janichsan]

Any further news on this following today’s announcement? Would the software Apple described enable an older macOS to run?

It was in the keynote, mentioned by name. They're calling this iteration 'Rosetta 2,' with that new version number, though. Some of the details under the hood are different. I don't know if any articles cover it yet.

Yeah, what I mean is, would Rosetta 2 enable an older macOS to run via a VM? Or would it only work to Translate software?

I was admittedly a bit surprised that they showed virtualisation, but I guess they wanted to alleviated some fears expressed by some of their users.

However...

On closer inspection, little about what I said really changed.

1. Rosetta 2 is not an (x64) emulator, it's a binary translator. As such, it is useless for VMs with Intel (or PPC) based operating systems. Generally, it translates the binaries before launch (useless for virtualisation), the mentioned on-the-fly translation for specific purposes most likely takes too much time to be useful for virtualisation even in theory, and last but not least, it almost for sure can only translate Mac binaries.

2. Note how they carefully avoided talking about what kind platforms they can virtualise. They only mentioned they could run Linux in a VM, but did not specify which Linux. What they were running was Debian (or a close derivative), which does have Arm-native versions. So, chances are that you can only virtualise Arm-compatible operating systems.

Bottom line is that I'm still 99.5% sure that you still will not be able to virtualise older, Intel-based versions of macOS on the upcoming Arm-Macs.

 

[cmaier]

Yeah, what I mean is, would Rosetta 2 enable an older macOS to run via a VM? Or would it only work to Translate software?

About the Rosetta translation environment | Apple Developer Documentation

Learn how Rosetta translates executables, and understand what Rosetta can’t translate.

developer.apple.com

 

[Spectrum]

What Can't Be Translated?
Rosetta can translate most Intel-based apps, including apps that contain just-in-time (JIT) compilers. However, Rosetta doesn’t translate the following executables:

• Kernel extensions
• Virtual Machine apps that virtualize x86_64 computer platforms

Rosetta translates all x86_64 instructions, but it doesn’t support the execution of some newer instruction sets and processor features, such as AVX, AVX2, and AVX512 vector instructions. If you include these newer instructions in your code, execute them only after verifying that they are available. For example, to determine if AVX512 vector instructions are available, use the sysctlbyname function to check the hw.optional.avx512f attribute.


---

Thanks for the link...This is going to be a problem for me then because I have some important research data for a piece of PowerPC software that was never rewritten for Intel. It thus needs 10.6.8 (original Rosetta) to run.

Isn't the ability to virtualise Intel OS rather an important feature to be losing?