”No benefit?” So you’d prefer that we still support the Apple ][, centronics ports, etc.? You don’t think there was a benefit for dropping support of all that? That’s an incredible position to take.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
ARM v9 Instruction Set Officially Announced.
- Thread starter Colstan
- Start date
- Sort by reaction score
”No benefit?” So you’d prefer that we still support the Apple ][, centronics ports, etc.? You don’t think there was a benefit for dropping support of all that? That’s an incredible position to take.
You're also wrong here, there's no promise WoA will ever be licensable, so there's no way I could use it for work for now. Expecting it will happen doesn't buy you a whole lot...
I don't really understand what you're saying there, but yes, absolutely no benefit for me and for where I work. And fwiw, we still have a few centronics printers too.
No, this is where you would be wrong, at least for many, many use cases. I've worked in corporate environments where time is money. Improvements that allow people to work more quickly, efficiently, catch or correct mistakes, etc. are definitely worth it. Buying new equipment or updating/adding new software often pales in comparison to the amount of money many people make in salary and benefits. This is especially true for creative professionals.
As for WoA support, as Apple has stated, that ball is in MS's court. There are pros and cons for MS. Apple has provided a way to run macOS x64 software for the near future. They're still selling Intel Macs, and will be for another year or so.
Perhaps fabbing old processors will be a nice side business for Intel, seeing how they're having trouble making modern ones.
Real improvements are different! Rewriting just because the platform changed, i.e., no improvements, just making it run, never has a positive ROI. And for getting money for improvements, you have to have some kind of case for that improvement, and that may even mean a new platform if the case is good enough.
On WoA, of course it's in Microsoft's court, never said it wasn't, but it is a problem for me running my stuff on the Apple platform. Remember I don't care about the hardware, I care about running the software I need...
Right, but gone are the days of just releasing a piece of software and then not touching it again. Unless it's a specific use case where it's running on a machine that isn't connected to the internet, at minimum you should expect bug fixes and security updates.
For Apple software, it's often as trivial as downloading the new version of Xcode, making some minor changes, then recompiling. Apple is usually pretty good about helping developers stay current because it is in their best interest to do so.
For Apple software, it's often as trivial as downloading the new version of Xcode, making some minor changes, then recompiling. Apple is usually pretty good about helping developers stay current because it is in their best interest to do so.
Actually, they're not gone, that's still the way a lot of corporate stuff works.
I don't directly expose any PC's (or PC servers) to the internet. (from outside) It's pretty much up to the OS maker and the AV software maker for updates on internal PC's.
It's never that trivial for real software systems when you're changing OS/hardware.
ARMv9 doesn't require a rewrite like ARMv8 did. That's one of the advantages.
I would guess 99.9% of software will just be able to recompile into ARMV9 without any changes.
I work on software, specifically for ARM architectures. You can bet on it.
What it means for ARMv9 Macs and ARMv8 Macs in the future? Dunno. But it's not at all what you're making it out to be.
What robco74 is saying is absolutely correct. As a developer who works on ARM systems I can vouch for it. You should go back and do some research.
That's .1% that don't.
I work on software too.
Oh, I agree that this v8 to v9 change isn't much, but given their transition to the M1 and how they did it, plus the getting rid off x86 compatibility of their own OS, I can't trust them at all. Given the attitude around here I should like that, but that's not why I own Mac's at home.
You mean to say 100%. ARMv9 still uses the same AArch64 as ARMv8, it just adds some stuff to it.
They didn't get rid of x86 compatibility? Or are you talking about 32-bit x86? Well, that was long overdue.
I understand where you are coming from, but I have to say I strongly disagree with your point of view. Legacy software support is a huge blocker of progress in both hardware and software innovation. One should not keep obsolete platforms around just because some people have been using the same software for 15 years and the developer doesn't feel like upgrading their codebase to modern standard. This unhealthy obsession with backwards compatibility reward bad work ethics, lazy software design, and makes the situation worse for everyone.
No, there is nothing about v9 that would encourage apple to drop support for v8 anytime sooner than it already would. The security differences are not a huge deal. I can’t promise you that you will have full OS support for 10 years - Apple does what it does. But my main computer is a 2016 Mac and I fully intend to buy whatever the next 16” MBP, with the expectation that I will be happy using it for 5 years, and it will be supported for at least a couple years more than that with OS updates so that I can retire it for use as a media server.
For me, it's about buying an expensive computer today and the concern of potentially having it dropped from OS support in less than 5 years. It's akin to people buying the first 32bit Core Duo Macs.
But if v9 is just an extra extension, then my worry is probably unwarranted. It's just that v9 is right in the corner.
The difference between 64-bit and 32-bit core duo macs is so much bigger a deal than the difference between v9 and v8 that it isn’t even funny. Apple may not even use most of the v9 extensions.
V9 vs. v8 is more like when Intel or AMD come up with a new chip with a couple extra instructions (Which happens all the time).
I guess different businesses have different priorities, but IMHO, businesses have to make tough decision sometimes to move on with times based on risk management.I don't really understand what you're saying there, but yes, absolutely no benefit for me and for where I work. And fwiw, we still have a few centronics printers too.
I know of an example where a company where they have a legacy bespoke software developed by a third party, which decided to move abandoning the support. The software is used to control certain equipment in the company's workflow. This has the potential of disrupting the company's operations should the software break down or the machine running the software breaks and replacements parts could no longer be bought.
You could say that Windows support backward compatibility, even for Windows 10/Server 2019, but the question is, will the business be 'brave' enough to run it's operations on unsupported software? I know most large corporations will not. Over time, small/medium size corporations will not be willing to pay the 'exorbitant' cost of support to keep legacy systems running.
IMHO, once Microsoft's cloud business is bigger than it's Windows business, we will start seeing Windows support getting cut, and the first thing to go will be legacy support.
At the end of the day, it's always risk vs cost benefit or whatever metric is important to businesses.
Well rather than worry about the new instruction set, perhaps worry that Apple will soon be making 128-bit CPUs and sunsetting 64-bit support.
Memory will have to get a hell of a lot cheaper before that happens
I don't think we're even close to exhausting the memory address space provided by a 64-bit CPU. I don't think we'll see anything larger than 64-bit in my lifetime.Well rather than worry about the new instruction set, perhaps worry that Apple will soon be making 128-bit CPUs and sunsetting 64-bit support.
And as @cmaier stated in an earlier post, memory has to get really, really cheap and density to shoot way up, for the industry to need to move to larger than 64-bit.
I can see the argument for larger register size tho. as it can probably speed up computation?
Does it make sense to have a CPU with 128-bit register with 64-bit addressing?
I don’t think so. In x86-64, our impetus for the 64-bit addressing was primarily for the address space, and we had a mode that did 32-bit calculations and 64-bit addressing.
I guess one thing you *could* do with 128-bit integers is eliminate the need to use floating point math for numbers that are small enough. Doing 128-bit math wouldn’t likely speed up most algorithms, because you are seldom dealing with numbers outside the 64-bit-representable range. Having 128bit registers that can hold multiple 64-bit values might be useful (much like x86 supports for its smaller formats). You could make every op potentially a SIMD op. But we already have SIMD, so…
That's a bad spot, I agree, but they're screwed if they need something different, whether they updated to the latest and greatest or not -- that can happen to any lifecycle in any software or hardware.
Yep, in spades, we already do. It's not at all uncommon!
Sure they will -- that's a heck of a lot cheaper than spending money on updating the software. Software costs are WAY more than hardware.
No way.
Absolutely.
So, ignoring the fact that these users have to go somewhere before Windows really enters a phase of cutbacks in resourcing, I think this is actually backwards.
The last thing to go would be the legacy support, as the folks who will keep using Windows are going to be folks with existing workflows that have to keep working (including legacy stuff), or enthusiasts (keep in mind that the PC is still about half the traditional gaming market) that benefit from compatibility with existing APIs.
Legacy support in Windows will die when Windows dies.
I agree, to a point. I’ve seen companies implode playing a game of chicken with the need to modernize. Which is not great when that company tied their destiny to these legacy platforms even as they withered due to age.
