I use my 2019 MacBook Pro for all my development these days. For things like Node.js or Java or Python and other high-level language and IDE stuff the switch to ARM probably won't make much difference.
But I also rely on a Parallels VM environment to run Windows and Linux VMs. My development target is sometimes those platforms and I need to bench test. Sometimes I have Windows-only applications like Solidworks. (Come to think of it what about the Microsoft Office suite? Are they on-board with migrati
I can imagine an emulation mode working but I can't imagine it would be very fast. I wouldn't buy a new system to go backwards in performance.
Apple has been putting a lot into LLVM over the last decade. Except for dynamic code execution, where a VM is running on top of the OS and generating the wrong instruction set on the fly, it will precompile the x86-64 into ARM64. My guess is the x86 is essentially treated like the source code of a particularly pedantic programming language, turned into an intermediate LLVM representation, then re-compiled down into ARM. It will be interesting to see how fast it is, but there's no particular reason why it would have to run as slow as the emulators of old.
It's amazing how many ignorant people pretend to be experts on/.
If you guess x86 object can be considered a "particularly pedantic programming language", whatever than means, stop guessing. You don't understand what is required.
Are you saying itâ(TM)s impossible to turn a binary back into an LLVM intermediate representation? There are already open source projects that do exactly that (see RetDec), and I assume with Apple willing to throw a billion dollars at the transition that theyâ(TM)ll do even better.
Itâ(TM)s easily possible that the reverse-x86 LLVM representation wonâ(TM)t be as easily optimized for ARM as one generated from original source, but I wouldnâ(TM)t bet on it being THAT much slower.
I highly doubt they are even attempting to recompile an entire x86_64 binary into ARM64, I think that's just a chinese whispers thing that's happened from the single statement we've heard. It's far more likely it's just plain old DynaRec and JIT just with the added trick of getting a head start on it before first run. Imagine if a Java program got to run the JIT compilation step on install instead of first run, people might have less of a "Java is slow" opinion if that had been the standard from the beginni
"An organization dries up if you don't challenge it with growth."
-- Mark Shepherd, former President and CEO of Texas Instruments
I'll be hanging on to x86 Mac for a while (Score:4, Interesting)
I use my 2019 MacBook Pro for all my development these days. For things like Node.js or Java or Python and other high-level language and IDE stuff the switch to ARM probably won't make much difference.
But I also rely on a Parallels VM environment to run Windows and Linux VMs. My development target is sometimes those platforms and I need to bench test. Sometimes I have Windows-only applications like Solidworks. (Come to think of it what about the Microsoft Office suite? Are they on-board with migrati
Re:I'll be hanging on to x86 Mac for a while (Score:2)
I can imagine an emulation mode working but I can't imagine it would be very fast. I wouldn't buy a new system to go backwards in performance.
Apple has been putting a lot into LLVM over the last decade. Except for dynamic code execution, where a VM is running on top of the OS and generating the wrong instruction set on the fly, it will precompile the x86-64 into ARM64. My guess is the x86 is essentially treated like the source code of a particularly pedantic programming language, turned into an intermediate LLVM representation, then re-compiled down into ARM. It will be interesting to see how fast it is, but there's no particular reason why it would have to run as slow as the emulators of old.
Re: (Score:1)
It's amazing how many ignorant people pretend to be experts on /.
If you guess x86 object can be considered a "particularly pedantic programming language", whatever than means, stop guessing. You don't understand what is required.
Re: I'll be hanging on to x86 Mac for a while (Score:2)
Are you saying itâ(TM)s impossible to turn a binary back into an LLVM intermediate representation? There are already open source projects that do exactly that (see RetDec), and I assume with Apple willing to throw a billion dollars at the transition that theyâ(TM)ll do even better.
Itâ(TM)s easily possible that the reverse-x86 LLVM representation wonâ(TM)t be as easily optimized for ARM as one generated from original source, but I wouldnâ(TM)t bet on it being THAT much slower.
And App
Re: (Score:2)
I highly doubt they are even attempting to recompile an entire x86_64 binary into ARM64, I think that's just a chinese whispers thing that's happened from the single statement we've heard. It's far more likely it's just plain old DynaRec and JIT just with the added trick of getting a head start on it before first run. Imagine if a Java program got to run the JIT compilation step on install instead of first run, people might have less of a "Java is slow" opinion if that had been the standard from the beginni