Language:
switch to room list switch to menu My folders
Go to page: First ... 14 15 16 17 [18] 19 20 21 22 ... Last
[#] Fri Feb 21 2020 07:29:06 EST from Ragnar Danneskjold

[Reply] [ReplyQuoted] [Headers] [Print]

I've been surprised how long the iPad's have lasted. I've got a few of them, but I'll be honest, I use them a lot less these days because the phones are larger and more convenient now.

[#] Tue Feb 25 2020 15:31:02 EST from LoanShark

[Reply] [ReplyQuoted] [Headers] [Print]


https://www.tomshardware.com/news/apple-mac-arm-cpus-processors-kuo

ARM processors in Macbooks?

This raises more questions than it answers. What happens to the Mac Pro? Will Apple continue with an Intel alternative for laptops?

Unfortunately this would probably mean the death knell for the Macbook as a usable development platform, at least for the short/medium term :-/

[#] Tue Feb 25 2020 15:31:42 EST from LoanShark

[Reply] [ReplyQuoted] [Headers] [Print]


Wouldn't be the first time some industry pundit speculated wrongly about an Apple product...

[#] Thu Feb 27 2020 20:41:28 EST from Ragnar Danneskjold

[Reply] [ReplyQuoted] [Headers] [Print]

I've been hearing that rumor for a few years.

My guess is they'll do something like Microsoft is doing with the new Surface Pro X and see how it works out....

[#] Thu Feb 27 2020 21:37:02 EST from IGnatius T Foobar

[Reply] [ReplyQuoted] [Headers] [Print]

I heard the ARM thing mentioned on a podcast I was listening to this morning, but it wasn't sourced. I find it hilarious that the writer of this article feels it is necessary to mention "Switching from Intel processors to custom Arm chips would also require changes to macOS." Heh. Ya think? It's not as if Mac users haven't had to deal with that twice already. Apple is good at that by now.

Ironically, a modern Macintosh has more in common with the original IBM PC than it does with the original Macintosh, as far as the hardware goes. They could build a machine that is just a Dell with a custom bootloader and a fancy case, and no one would know the difference.

MacOS on ARM ... probably inevitable, unless Apple learns to Think Different and starts building machines with AMD processors. They probably would have a lot more leverage with AMD than with Intel, and they could probably strongarm (heh) AMD's supply chain into ramping up manufacturing when they need to.

Merging the macOS and iOS operating systems is certainly possible, but Apple probably wants everyone to buy two machines instead. But it would be hilarious to see them get it right, just to show Microsoft how it's done, years after we're *still* wanting to see the remnants of Windows Phone removed from Windows 10.

[#] Fri Feb 28 2020 13:41:41 EST from LoanShark

[Reply] [ReplyQuoted] [Headers] [Print]


According to the one article, Apple confirmed those plans. But who knows if that's true or what it means exactly.

ARM doesn't have the single-thread performance for a lot of use-cases

[#] Thu Apr 23 2020 15:51:03 EDT from IGnatius T Foobar

[Reply] [ReplyQuoted] [Headers] [Print]

The rumors seem to be getting louder. Anti-Big-Gulp News claims [https://tinyurl.com/y7znoctc] that the fruity bunch have three SoC components in the pipeline based on the A14 ARM processor, which will also appear in the next iPhone. The project, code named "Kalamata" (which is an olive, not an apple) is rumored to be based on a 5-nanometer fab, also in line with what will be in the next iPhone.

The article acknowledges that low-end laptops are the entry point because of single thread performance.

If the Anti-Big-Gulp paywall gives you a problem, a secondhand-account but more readable article is at The Register [https://www.theregister.co.uk/2020/04/23/apple_arm_chips/].

IG Spec: I wonder if it would be viable to place Intel *and* ARM processors in the same computer, similar to how the Commodore 128 had both 6502 and Z-80 onboard. I imagine it would require some sort of special-purpose hypervisor to keep everything running in symphony.

[#] Thu Apr 23 2020 18:50:07 EDT from LoanShark

[Reply] [ReplyQuoted] [Headers] [Print]


I'm surprised at what they snuck in there at the end--Apple inks a short-term deal with Qualcomm for modems--and then buys Intel's entire modem business the next month.

hmmmmmm

[#] Mon Apr 27 2020 20:34:57 EDT from Ragnar Danneskjold

[Reply] [ReplyQuoted] [Headers] [Print]

I'm going to guess it will be in a laptop, kind of like Microsoft's Surface X Pro, which runs their SQ1 processor.

I think for a lot of people, they don't need much. Email, a browser, a word processor and a handful of apps that need to run well. Lots of battery life, cellular modem.



[#] Tue Apr 28 2020 09:31:42 EDT from IGnatius T Foobar

[Reply] [ReplyQuoted] [Headers] [Print]

That's the easy part. The more interesting speculation is over what they will do to Mac OS to make it run on ARM. Apple's ability (and willingness) to transition the OS to different CPU architectures is well known and they are pretty good at it -- hybrid binaries, transcoding, etc. This would signal that they're eventually going to move everything.

Microsoft's ARM efforts have largely been a failure because they used a version of Windows that was identical to, but incompatible with, the mainline AMD64 version of Windows. That works fine with Linux when there's a universe full of existing source code repositories that can simply be rebuilt, but when the *primary* value of your OS is compatibility with the last few decades of off-the-shelf binaries, an incompatible version is not going to sell.

So yes, Apple does this better. But what if Apple really does want to build an incompatible device here? It may not make sense to run Mac OS at all. In that case they could call it a Mac, but it could be more prudent to have it run a version of iPad OS enhanced with a window manager.

[#] Tue Apr 28 2020 10:45:48 EDT from LoanShark

[Reply] [ReplyQuoted] [Headers] [Print]


iOS is Darwin, so they already did most of the porting work. They've had that multarch binary thing in place ever since NeXTstep. Surely they'll just keep doing it the way they did for the PPC/i386 and i386/amd64 arch transitions. It'd be surprising to see x86 emulation work well, though.

[#] Tue Apr 28 2020 16:18:55 EDT from IGnatius T Foobar

[Reply] [ReplyQuoted] [Headers] [Print]

I knew that they used Universal Binaries to manage the transition from PPC to i386, but I didn't know they used them to manage the transition from i386 to AMD64. That's actually quite a clever way to manage the transition to 64-bit without keeping a separate set of 32-bit binaries around forever.

I have a hard time seeing Universal Binaries as a semi-permanent solution rather than a transition, though. Even if the build tools abstract away as much of the cross-compiling as possible, you'd still have to test on every architecture. Perhaps they have a "fast" ARM in the works.

[#] Fri Jun 26 2020 09:24:25 EDT from IGnatius T Foobar

[Reply] [ReplyQuoted] [Headers] [Print]


Well, they gone and done it.

[ https://www.apple.com/newsroom/2020/06/apple-announces-mac-transition-to-apple-silicon/ ]

The first ARM-based Mac will be released this year, and they expect to take about two years to complete the transition. Meanwhile they will continue to release new versions of Mac OS for x86 for some time.

The press release says that they plan to make "iOS and iPadOS apps available on the Mac without any modifications." There is already some speculation that Apple may try to force-march Mac OS developers to the App Store by making this the point where they prohibit sideloading.

Although I'm not an Apple fan, I've gotta give them props for this -- they have a habit of making moves that anyone else would say "the industry isn't ready for this yet" and succeeding.

[#] Sat Jun 27 2020 14:35:58 EDT from LoanShark

[Reply] [ReplyQuoted] [Headers] [Print]


It's already become problematic for developers, one of the most recent versions of MacOS finally prohibited unsigned binaries, I believe.

MacOS will soon be dead-on-arrival as a serious development platform; many compilers out there require some serious single-thread performance that ARM will continue to lack for the foreseeable future.

[#] Sat Jun 27 2020 17:18:10 EDT from Ragnar Danneskjold

[Reply] [ReplyQuoted] [Headers] [Print]

It will bitch about unsigned binaries, but you can force them to run.

[#] Sat Jun 27 2020 17:33:24 EDT from IGnatius T Foobar

[Reply] [ReplyQuoted] [Headers] [Print]

I'm ok with the warning, as it permits you to dig into your settings and say "I know what I'm doing" while giving Joe Sixpack an easy way to stay inside the garden and not worry about it. The same could be said for sideloading apps onto an Android device.

We all know that Apple, Google, and Microsoft would *love* to restrict their platforms to 100% app store programs only. They just haven't figured out how they can get away with it, without a user revolt. I think the exception is iOS which has always been that way. Microsoft tried it with Windoze 10S and nobody bought it.

I have to assume that Apple's "two year plan" must involve "making ARM go fast" in single-threaded workloads. Maybe they have a chip in the works that combines fast, power-hungry cores with slow, power-sipping cores. Qualcomm had that a number of years ago, but the "Ofast" cores were the current state-of-the-art ARM, not something as fast as the x86 watt-guzzlers of the day.

It is already being reported that Apple is going to spend a lot of effort encouraging and assisting developers to optimize for multithreaded performance.
I am skeptical that this alone can make current-generation ARM competitive with Intel on a mainline Mac desktop.

AMD should build a Ryzen chip with microcode to run ARM instructions, to bail Apple out when they fail :)

[#] Sat Jun 27 2020 17:41:55 EDT from LoanShark

[Reply] [ReplyQuoted] [Headers] [Print]

2020-06-27 17:18 from Ragnar Danneskjold
It will bitch about unsigned binaries, but you can force them to run.


I mean, that's been the case for the last 10 years. I don't run MacOS anymore so I can't personally confirm, but apparently it became impossible to even force them?

[#] Sat Jun 27 2020 17:45:41 EDT from LoanShark

[Reply] [ReplyQuoted] [Headers] [Print]

It is already being reported that Apple is going to spend a lot of
effort encouraging and assisting developers to optimize for
multithreaded performance.

That'll never happen to a sufficient degree. Some software is just stuck in single-thread land for the foreseeable future. The Java compiler is one example. It's got such a long history and they have to focus on correctness over performance. Yeah, I think a lot of end-user-facing workloads will get better about multithreading. But developer-facing stuff is not in that category, and neither is some of the workstation stuff that's out there.

[#] Sat Jun 27 2020 17:46:57 EDT from LoanShark

[Reply] [ReplyQuoted] [Headers] [Print]

AMD should build a Ryzen chip with microcode to run ARM instructions,

to bail Apple out when they fail :)

Funny, I'm pretty sure AMD's embedded management coprocessor is an ARM chip. But of course you can't run user-programmable code on that, it's a walled garden.

[#] Tue Jun 30 2020 12:12:41 EDT from IGnatius T Foobar

[Reply] [ReplyQuoted] [Headers] [Print]

I guess we will see if it's possible to make an ARM core "go fast". There was a company about 10 years ago that was building a machine that accelerated Java workloads using a proprietary massively-parallel machine. To make it work you had to use a special JVM that sent your conpute over to that machine.
They had teams of consultants to help customers get started, and most of what they did was help customers parallelize their workloads.

Unsurprisingly, the business model failed.

Go to page: First ... 14 15 16 17 [18] 19 20 21 22 ... Last