Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
i had passed mine to a kid, it only runs plugged it. I know they used it for something as late as last year. It was a great purchase, and my move back to Apple (My first Mac being an SE back in the 80s).

Up until about a year ago I was using an old 2006 MacBook Pro (admittedly Intel-based) as our home media server. The few times I actually sat down and used it for something else, the performance was surprisingly good.

For a lot of day-to-day uses, processor speed essentially hasn't mattered for 15-20 years.
 
a curiosity at best.
Really.... Who's still using a PPC Mac? Sure, there's lots of people with old PPC macs stuffed somewhere in a basement under that pile of ZIP drives- but actually using one?

I can’t see anyone using an old PPC Mac as a daily use machine. However, PPC and ARM are the same type of architecture: RISC. RISC-V is growing popularity as alternative to ARM, where it doesn’t require holders to buy the spec. Samsung and NVIDIA are holders of RISC-V, the latter is moving entirely away from the ARM spec, but it’s the same architecture and very similar to it. Apple and ARM have been in deep with one another for quite some time, there’s a chance they could buy it out.

Last year, IBM turned PowerPC into open source technology, so there’s a chance someone could grab it and commercially use it as well.

There‘s a big storm of development brewing in RISC, largely in part to mobile devices.
 
I can’t see anyone using an old PPC Mac as a daily use machine. However, PPC and ARM are the same type of architecture: RISC. RISC-V is growing popularity as alternative to ARM, where it doesn’t require holders to buy the spec. Samsung and NVIDIA are holders of RISC-V, the latter is moving entirely away from the ARM spec, but it’s the same architecture and very similar to it. Apple and ARM have been in deep with one another for quite some time, there’s a chance they could buy it out.

Last year, IBM turned PowerPC into open source technology, so there’s a chance someone could grab it and commercially use it as well.

There‘s a big storm of development brewing in RISC, largely in part to mobile devices.
I don’t want to sound like I’m picking your message apart. “What use is there for old PPC Macs anymore” that the poster asked is very different than “what use is there for RISC anymore” that you seemed to be answering. Are you suggesting that there is or could be a use case for PPC Macs because of the developments in RISC? If so, I’d be curious to hear more as I also have a basement full of G4s and G5s (and a pile of Zip and Jaz drives 😛)
 
  • Like
Reactions: progx
Instead you need to build the PPC binary on an old version of Xcode running on an old version of OS X (in a VM or using separate boot disk or partition on an older Mac). That old Xcode will include the old proper libraries and frameworks. Then copy the binary to the current version of Xcode on a current Mac running a current macOS, and lipo your fat binary together, before code-signing or notarization.
I remember doing basically this for an iOS app I used to make, back when Apple dropped support for armv6. I still had users on older devices, and there was no technical reason why they couldn't still be supported.
 
Outside of tinkerers, few will try to target ppc these days. it’s also useless if the application uses newer frameworks that don’t have any ppc support.
 
Have several web servers running G5 and G4 applications. Although we're starting to see the units starting to fail from dried out capacitors and having to reapply CPU cooking paste.
usually you can get them for next to nothing but still have a lot of performance.

Last year I finally virtualized my Xserve G5, it was worth it for the power savings alone, the G5s are hot, power hungry machines. The substantially faster disk, ram, and processor on the virtual host makes up for the Rosetta performance penalty.
 
  • Like
Reactions: pmm7002641 and wtn
I'm not surprised.
I'd be more surprised if they could get a binary and assets working on OpenStep 4 PA-RISC, SPARC, Motorola 68k, PowerPC Macs, Intel 32-and-64-bit, and Apple Silicon.
I'd be even more surprised if they got OpenStep i386 and Mac OS X i386 working in the same binary!

Ostensibly you could link in NeXTStep 3.3 slices, but you would have to write all of the code in a way that supports manual retain-release unless you want to port the GNUstep libobjc2 runtime, and you'd also be limited to old APIs (assuming you build each slice from the same source).

I don't think you could realistically have i386 binaries working in macOS and OpenStep, because AFAIK they use the same magic number (unless enough of the basic runtime libraries happen to be in the same place so that you could write your own code that calls dlopen and handles the errors manually, in which case maybe...).

That said, any x86 Mac will run the x86-64 slice preferentially by default, so you would probably be better off just making the i386 slice be for OpenStep; as long as somebody doesn't manually tell it to open in 32-bit mode, it would theoretically work everywhere. :D
 
  • Like
Reactions: RandomDSdevel
Have several web servers running G5 and G4 applications. Although we're starting to see the units starting to fail from dried out capacitors and having to reapply CPU cooking paste.
usually you can get them for next to nothing but still have a lot of performance.
I think used Mac Pros are better $/performance. Energy usage is also a consideration.
 
  • Like
Reactions: JosephAW
Last year I finally virtualized my Xserve G5, it was worth it for the power savings alone, the G5s are hot, power hungry machines. The substantially faster disk, ram, and processor on the virtual host makes up for the Rosetta performance penalty.
Fortunately I don't have to pay for power at my site, just internet. Also acquired several G5 units for $20. Mac Pro are a little bit more.
 
  • Like
Reactions: kwikdeth
Until ARM becomes popular with business don't bet the farm about intel going away. I have no love for Fat Binaries.
Odds arm the ARM based mac will be screaming fast and if Apple has done their homework but then, they'll have the packages businesses run on their macs ready for the new architecture to run natively. Even then, their rosetta2 seems to be quite capable even on what's still merely a clocked down iPad Pro CPU. So anythign they release is likely to either have extreme long battery life, extreme light weight, or so much more powerful than any currentm, that everybody will forget about intel CPUs.
IfApple can;t achieve that, they'll not release it: it has to be an instant convincer to succeed.
[automerge]1594524090[/automerge]
Good stuff! Betting/hoeing VMWare & Parallels are hard at work.

And welcome back to the surface, Arnold!
VMware and Parallels do NOT emulate a different CPU.
They use the virtualisation features modern Intel CPUs have to do the "stuff" they do.

Rosetta (and Rosetta 2) emulate (or translate) a binary executable from one architecture to the other. There's a (big) cost to pay to do that (although Apple's rosetta 2 (intel code emulated on an ARM) seems to only use ~30% overhead, which is amazing in itself) The old Rosetta (PPC code on intel emulation) was not very efficient from wat I remember. My first use of it (with MSFT office e.g.) was a miserable experience as my old G4s did a fater job than mymuch nwere intel based CPUs emulating the PPC.
 
Last edited:
And folks out there, in the modern world, are telling us that iTunes is dead and we should just move on and adapt to the Music app, cowpokes.
 
  • Like
Reactions: RandomDSdevel
Odds arm the ARM based mac will be screaming fast and if Apple has done their homework but then, they'll have the packages businesses run on their macs ready for the new architecture to run natively. Even then, their rosetta2 seems to be quite capable even on what's still merely a clocked down iPad Pro CPU. So anythign they release is likely to either have extreme long battery life, extreme light weight, or so much more powerful than any currentm, that everybody will forget about intel CPUs.
IfApple can;t achieve that, they'll not release it: it has to be an instant convincer to succeed.
RE: Then they have the packages businesses run on their macs ready . . . Do you think Apple makes most business software that businesses use? Nope.
RE:So anything they release is likely to either have extreme long battery life, extreme light weight, or so much more powerful than any currentm, that everybody will forget about intel CPUs. . . . This sounds like it originated from a iPad fan, desktops and workstation don't work off batteries, either a laptop or desktop won't worry about small file size (lightness), and I have yet to see a program made so much more powerful just because of RISC code efficiency. o_O
 
  • Like
Reactions: Detektiv-Pinky
May I ask why someone would want to do this? Aren’t all PowerPC processors toaster speed at this point?

I will admit that from a technical standpoint it is neat.
 
It would be good, and while "anything is possible, in a real world, who would you even want to ? I can't imagine the amount of troubleshooting through code you'd be up against trying to see which architecture a bit of code that broke on x platform.It's easier on a universal app iPad/Mac I'd imagine but all 4 ? Particularly when it works ok on one, but not on another.
 
I am a retro gamer, Apples transitions are a pain in the butt. During the transition to intell, I lost AVP2, Diablo 2 etc. Now, they dropped support for 32bit apps, which means most "made for mac games" don't work, like Black & white 2, Diablo 2 (yet again after they patched it for intel), and all my old Dosbox games running on Boxer won't run anymore. Now we are going to ARM and alot of Steam games probably won't work. I have literally stayed on Mojave because of so many 32 bit apps I am running. It is super annoying!

If anybody can get the Dosbox app Boxer running (or know of an alternative) on the new macs I will be happy. Messaged the developer no luck :( Why couldn't apple just do a 32-bit translation layer for legacy apps?

 
It would be good, and while "anything is possible, in a real world, who would you even want to ? I can't imagine the amount of troubleshooting through code you'd be up against trying to see which architecture a bit of code that broke on x platform.It's easier on a universal app iPad/Mac I'd imagine but all 4 ? Particularly when it works ok on one, but not on another.
I think just from a technical point of view and to demonstrate powerful nature of macOS as a platform through its lineage and the ability to invest it through the various transitions. So, if you made a bet on Mac OS 10.0 in March 2001, just think of how cool, 20 years later, that app you wrote, can easily be installed on macOS 11 in March 2021.
 
but_why.gif
May I ask why someone would want to do this? Aren’t all PowerPC processors toaster speed at this point?
It would be good, and while "anything is possible, in a real world, who would you even want to ?

You guys are all no fun! :(

It's neat god damnit! I love the idea that you could copy a single .app bundle between three different Macs running completely different CPU architectures, from completely different eras, and it would still work on all of them without emulation.

I just hope someone does it, once ARM Macs are actually available. We need a good candidate. FFMPEG is a good one—a legitimately useful piece of software that can already be compiled for PPC, Intel, and ARM versions of Darwin—but it's a command line utility. I'd rather see a real app with a UI! :)
 
Last edited:
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.