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.
Mac OS X 10.4 (Tiger) Coming for the PC?
- Thread starter MacRumors
- Start date
- Sort by reaction score
OS X will never go to the x86 side as a product you can install on any computer you wish.
I think the cool thing in this article is the emulator possibilities that Apple could use in Mac OS for the PPC. What if we had a faster, more integrated x86 emulator built into the OS? Sure, maybe games would still be a bit laggy, but any other x86 program could run on the Mac? Basically, with the ability to run a lot of linux software already, and the ability to run most windows software as well, the sky would be the limit for Apple.
Of course, this is nothing but speculation.
I think the cool thing in this article is the emulator possibilities that Apple could use in Mac OS for the PPC. What if we had a faster, more integrated x86 emulator built into the OS? Sure, maybe games would still be a bit laggy, but any other x86 program could run on the Mac? Basically, with the ability to run a lot of linux software already, and the ability to run most windows software as well, the sky would be the limit for Apple.
Of course, this is nothing but speculation.
On IT morons:
And that would reallly show how much of a moron your IT guy is, since what exactly is "real" UNIX.
Mac OS 10 is NOT simply based on UNIX (and I hate the fact that Apple themselves use this phrasing "based", perhaps in an effort to not have to pay some licensing fees, etc, not sure).
Lets get it straight once and for all,
Mac OS 10 IS BSD 4.4 UNIX, freakin' period. (of course with some incredible, unparalleled and unmatched architectures above it, let's see Linux come up with such architectures and layers as Core Audio / Core Image, etc.
And in many cases (BSD 4.4 UNIX) is the choice over other UNIXes especially Linux: government, financial institutions, etc. for example.
And that would reallly show how much of a moron your IT guy is, since what exactly is "real" UNIX.
Mac OS 10 is NOT simply based on UNIX (and I hate the fact that Apple themselves use this phrasing "based", perhaps in an effort to not have to pay some licensing fees, etc, not sure).
Lets get it straight once and for all,
Mac OS 10 IS BSD 4.4 UNIX, freakin' period. (of course with some incredible, unparalleled and unmatched architectures above it, let's see Linux come up with such architectures and layers as Core Audio / Core Image, etc.
And in many cases (BSD 4.4 UNIX) is the choice over other UNIXes especially Linux: government, financial institutions, etc. for example.
SiliconAddict said:
*coughs*No single processor tower with multiple optical drives, desktop replacement notebook, or even a lowly card reader*coughs*
All I can say is this, if Google's rumored operating system as stable, virus free, innovative, and easy to use as OSX Apple could be in real trouble.
this could be alright. it would give creative window-user folk something to change to without spending a whole load of money on new hardware.
however i remain slightly skeptical. surely x86 hardware cant run PPC completely? again isnt this just an emulator?
would x86 OSX run slowler with less effects than its PPC brothers?
what about security? viruses and the likes?
on one hand i would love to transform my entire house network to OSX instead of my PB being the only one. but on the other hand surely OSX will not run as well on x86. surely? and then OSX will have to stock up massively on its driver database to run the plethora of 3rd party hardware available to PCs.
but i suppose Mac have been caught sleeping in the PCs bedroom recently (dropping of Firewire, win iTunes etc). who knows?
however i remain slightly skeptical. surely x86 hardware cant run PPC completely? again isnt this just an emulator?
would x86 OSX run slowler with less effects than its PPC brothers?
what about security? viruses and the likes?
on one hand i would love to transform my entire house network to OSX instead of my PB being the only one. but on the other hand surely OSX will not run as well on x86. surely? and then OSX will have to stock up massively on its driver database to run the plethora of 3rd party hardware available to PCs.
but i suppose Mac have been caught sleeping in the PCs bedroom recently (dropping of Firewire, win iTunes etc). who knows?
Transitive's technology wouldn't be used to run the OS in emulated mode.
The OS would be compiled for x86, and then this technology would be used to run applications without needing a recompile. That way, the amount of "emulated" code is kept minimal, and all system-level tasks are native x86 code.
Note: I'm not saying this is a good idea for Apple.
The OS would be compiled for x86, and then this technology would be used to run applications without needing a recompile. That way, the amount of "emulated" code is kept minimal, and all system-level tasks are native x86 code.
Note: I'm not saying this is a good idea for Apple.
wHo_tHe said:
Just read through this whole thread waiting for your comment exactly, wHo_tHe. Well said. This would make the most sense for Apple if they 'HP OS X business box' direction discussed above.
Just to play devil's advocate, why wouldn't Apple be interested in creating an "ecosystem" similar to what Microsoft has done with Windows? Apple could help drive hardware direction in the Intel world just MS has been doing for years.
But Apple would succeed at the cross-(processor)-platform game where Microsoft failed. Windows' (and, by extension, Windows applications') continued reliance on Intel and AMD for forward compatibility points to the weakness of their engineering approaches and OS architecture.
Apple gives select OEMs an alternative to Windows hell while keeping IBM's PPC goodies for themselves. Sure, why not?
That said, creating an alternative to Virtual PC (also mentioned above) seems a possibility for this reason: the ability to boot x86 OSes on a PPC box might be helpful getting Macs in the door of many businesses.
It's been said that future dual-core PPCs will support IBM's 'virtualization technology'. I don't know exactly what that means, but it's supposed to support running multiple operating systems simultaneously. Does this mean we'll be able to dedicate one of our cores to running that 'legacy compatibility environment' known as "Windows"? Heheh, sweet.
That's a great point you mention, dglow. Would Apple be interested in this idea? Who knows...dglow said:
As far as the future of PPC goes, the 'virtualization technology' you speak of is present in the Power5 CPU from IBM. I'm not sure if it will make it into PowerPC, though, since Apple may not have a use for such a feature. Also, even if the technology makes it into a future PowerPC processor, you still can't run Windows without being able to somehow execute x86 code - as far as I know, this technology doesn't solve that problem.
Remember the Clones?
Apple has already licensed it's OS before....The PowerPC Clones.
They ate into Apple's profit and thankful someone (SJ) killed them off before Apple went under.
Why go down that road again?
Apple has already licensed it's OS before....The PowerPC Clones.
They ate into Apple's profit and thankful someone (SJ) killed them off before Apple went under.
Why go down that road again?
There is absolutely no reason for Apple to use anything from Transitive. The OS already supports multiple target architectures. It has since before it was an Apple OS. This is nothing new. NeXTStep supported 4 architectures out of the box. Applications on it, granted that the programmer checked the boxes for the correct targets, would run on multiple architectures under NeXTStep at full speed with no issues.
Since this is all based on NeXT Technology, the old Apple stuff, (carbon included) does not benefit from this. This means virtually all commercial software would have to be re-written for cocoa. The reason for this is that Cocoa apps are "application bundles" which are in fact folders named xxxx.app. The actual executable code is located in xxxx.app/Contents/MacOS. Apple already uses GCC to compile code. GCC supports multiple architectures. They'd have to add hooks for x86 compiles, and make sure all of their libraries were ported, which I think is mostly there (see Darwin x86) however poor the driver support is.
All that Apple would have to do beyond that is to make Xcode also create a xxxx.app/Contents/x86 folder containing the x86 binary and add a small hook in OSX to run the appropriate binary to the architecture it was running on. Since this has already been done in the past with NeXTStep, and because NeXT was designed this way, this is not that big of a deal. BTW the binary is actually rather small, so application sizes would not expand that much. Example: Aspyr Game Agent 1.1 2.5M, binary portion 248k. Pages: 395M, binary portion 2.4M. Get the picture?
Now here's the deal killer: Drivers. Millions of questionably designed pieces of hardware that _mostly_ follow a standard (or not), that exist in the windows world from companies who are not going to go to the trouble of writing a driver no matter how easy it is. Why M$ could yank their "designed for Windows XP" certification or make it cost more. And even if M$ doesn't resort to this kind of BS (and it has in the past), companies would be stupid to risk it.
Also with regards to the IT guys and UNIX. I am a serious IT guy. I manage networks, servers (Windows, Linux, OSX at work) & (BSD, Solaris at home for fun). I have 8 Operating systems running in my house across 5 architectures. OSX is BSD 4.4 with a Mach micro-kernel. I imagine your "UNIX" IT guys would choke if faced with Digital Unix (Now called True 64), AIX, or Solaris. You might ask them what /dev/wd0s1 refers to.... it's actually an evil question, since FreeBSD has moved to a different naming convention, they now use /dev/ad0s1... (refers to first IDE drive, first partition, the equivalent of C: in the windows world.)
Anyways 90% of what I used to rely on Slackware and FreeBSD for I now use OSX for and I get M$ office and some Adobe stuff with a shared clipboard to boot. If you have source, most of the libraries have been ported, and you can mostly just compile and go. OSX is a strange UNIX, but then again every flavor has it's oddities. I've not heard anyone trash IRIX for not being UNIX, and it's certainly very odd in many respects.
Important note: At home, I still use Slackware and FreeBSD on x86 for servers because it is free, and since I use no GUI on it, very lightweight (memory and CPU).
Since this is all based on NeXT Technology, the old Apple stuff, (carbon included) does not benefit from this. This means virtually all commercial software would have to be re-written for cocoa. The reason for this is that Cocoa apps are "application bundles" which are in fact folders named xxxx.app. The actual executable code is located in xxxx.app/Contents/MacOS. Apple already uses GCC to compile code. GCC supports multiple architectures. They'd have to add hooks for x86 compiles, and make sure all of their libraries were ported, which I think is mostly there (see Darwin x86) however poor the driver support is.
All that Apple would have to do beyond that is to make Xcode also create a xxxx.app/Contents/x86 folder containing the x86 binary and add a small hook in OSX to run the appropriate binary to the architecture it was running on. Since this has already been done in the past with NeXTStep, and because NeXT was designed this way, this is not that big of a deal. BTW the binary is actually rather small, so application sizes would not expand that much. Example: Aspyr Game Agent 1.1 2.5M, binary portion 248k. Pages: 395M, binary portion 2.4M. Get the picture?
Now here's the deal killer: Drivers. Millions of questionably designed pieces of hardware that _mostly_ follow a standard (or not), that exist in the windows world from companies who are not going to go to the trouble of writing a driver no matter how easy it is. Why M$ could yank their "designed for Windows XP" certification or make it cost more. And even if M$ doesn't resort to this kind of BS (and it has in the past), companies would be stupid to risk it.
Also with regards to the IT guys and UNIX. I am a serious IT guy. I manage networks, servers (Windows, Linux, OSX at work) & (BSD, Solaris at home for fun). I have 8 Operating systems running in my house across 5 architectures. OSX is BSD 4.4 with a Mach micro-kernel. I imagine your "UNIX" IT guys would choke if faced with Digital Unix (Now called True 64), AIX, or Solaris. You might ask them what /dev/wd0s1 refers to.... it's actually an evil question, since FreeBSD has moved to a different naming convention, they now use /dev/ad0s1... (refers to first IDE drive, first partition, the equivalent of C: in the windows world.)
Anyways 90% of what I used to rely on Slackware and FreeBSD for I now use OSX for and I get M$ office and some Adobe stuff with a shared clipboard to boot. If you have source, most of the libraries have been ported, and you can mostly just compile and go. OSX is a strange UNIX, but then again every flavor has it's oddities. I've not heard anyone trash IRIX for not being UNIX, and it's certainly very odd in many respects.
Important note: At home, I still use Slackware and FreeBSD on x86 for servers because it is free, and since I use no GUI on it, very lightweight (memory and CPU).
smurfjammer said:
1. They licensed the OS to small unknown companies and denied licenses to the major computer makers like Gateway, the kind with a marketing budget and an established user base.
2. During the clone era, Apple made a run of the mill product that was inferior to the clones. These days The iMac destroys similar AIOs like the Gateway Profile 5 when it comes to price. A barebones profile with a bottom of the line celeron costs as much as the 1.8 with the super drive. Then there's the Mac mini. Nobody has seen anything like it. But what if you want a MicroATX-type tower or a desktop replacement notebook? You're forced to choose between hardware that suits your needs and a good operating system. Many choose to brave windows.
10.4 for PC, good and bad
People love to have a system with virus free. So if OSX is license for PC platform, it will be a huge market.
But the risk is there. If 10.4 performance in base PC is better than Apple platform, Apple will loose his hardware business market. If in the other way round, Apple actually give a very bad impression in OS performance to the new market.
I like Mac cos both software and hardware are design by it's own company. Make sense that it's design and debug using the standard platform. Unlike PC, it's design base on one platform and sell out with requirement compatible platform. It would be like another MS if really happen.
People love to have a system with virus free. So if OSX is license for PC platform, it will be a huge market.
But the risk is there. If 10.4 performance in base PC is better than Apple platform, Apple will loose his hardware business market. If in the other way round, Apple actually give a very bad impression in OS performance to the new market.
I like Mac cos both software and hardware are design by it's own company. Make sense that it's design and debug using the standard platform. Unlike PC, it's design base on one platform and sell out with requirement compatible platform. It would be like another MS if really happen.
Jinorasa said:
Not gonna happen. Even IF we believe the hype that this new emulation software is 80% efficiency (unlikely; the fastest PowerPC emulators out there currently are 10% of processor speed under best conditions with an optimized build, and usually less, not to mention no Quartz Extreme support)...there would be a 20% drop in processor speed. Meaning that unless Mac processors are waaaaay slower than x86, the x86 will slow down enough because of emulation that the Mac will be a bit faster.
I have an emulation question:GFLPraxis said:
Anyone know how efficient the best x86 emulators for PowerPC are? I remember reading on the Internet that the reason it took so long for PearPC to appear is because emulating x86 on PPC is considerably easier than PPC on x86 due to architectural differences (PPC has more non-rename registers than x86, for example).
Why emulate?
Why would you emulate something like a CPU when you can use an OSes architecture to have multi-architecture native code? This is one of the things that makes UNIX so cool. If you have the same libraries and OS across a plethora of hardware, you can just recompile for _native_ speed.... You can run the same version of Apache, on PPC, x86, PA-RISC, Alpha, MIPS, ARM, Sparc, Power, or whatever as long as the OS and libraries match, provided you recompile. It is stupid in this case to emulate. The reason Java caught on (emulated) is that windows is not compatible with anything else (sometimes even other versions of windows), and it allows systems that don't have an architecture that allows for multiple architecture apps/installers (Redhat uses SRPMS which can be installed on any redhat platform on virtually any supported hardware), to have some of the benefits of write once, run anywhere.
wrldwzrd89 said:
Why would you emulate something like a CPU when you can use an OSes architecture to have multi-architecture native code? This is one of the things that makes UNIX so cool. If you have the same libraries and OS across a plethora of hardware, you can just recompile for _native_ speed.... You can run the same version of Apache, on PPC, x86, PA-RISC, Alpha, MIPS, ARM, Sparc, Power, or whatever as long as the OS and libraries match, provided you recompile. It is stupid in this case to emulate. The reason Java caught on (emulated) is that windows is not compatible with anything else (sometimes even other versions of windows), and it allows systems that don't have an architecture that allows for multiple architecture apps/installers (Redhat uses SRPMS which can be installed on any redhat platform on virtually any supported hardware), to have some of the benefits of write once, run anywhere.
Yes, I agree with you that UNIX makes the concept of "write once, run anywhere" relatively simple - usually a painless recompile is all that's needed. However, you avoided my question rather than answering itMagnusDredd said:
motuman said:
Since im going to be pedantic here.
A real UNIX would be any OS derived from System V from Bell Labs then sold to Univel then sold to Novell then sold to SCO.
BSD was a seperately derived "UNIX" developed at Berkley, alot of lawsuits surrounded this OS... but have long been settled.
Apple doent have to worry about the phrasology of calling OS X as being based on UNIX.
motuman said:
YES Lets get it straight once and for all.
OS X IS NOT BSD 4.4 UNIX.
In the "UNIX" world the kernel IS the OS. All the other stuff is basically just that... other stuff.
OS X .....DOES NOT... and I repeat DOES NOT... use BSD's kernel.
They use a hodgepodge kernel "BASED ON" BSD and MACH with IO KIT thrown in and they call their hodgepodge kernel DARWIN.
motuman said:
I take greatest exception to this....
Linux is fastly becoming the number one "UNIX" type OS PERIOD.
Almost ALL major vendors are replacing their UNIX's with Linux.
For example
IBM and SGI are converting everything to LINUX
and SUN will follow soon.(as Soon as they see their recent opensourcing of Solaris doesnt help)
I don't know if His Steveness ever personally demoed Openstep on the PCs it ran on (and Windows it ran under) but he did make an OS (or at least, it's frameworks) for the PC, and lest we forget, Darwin still runs on PC hardware, so in a way, Mac OS X runs on PC hardware for free.Phillip said:
Mac-Xpert said:
yeah, and software houses stop porting and writing OsX-native applications and kill the Mac platform off... what a brilliant idea
jeez
vSpacken
vollspacken said:yeah, and software houses stop porting and writing OsX-native applications and kill the Mac platform off... what a brilliant idea
jeez
vSpacken
Windows is an OS that runs on the x86 platform, not the platform itself. The game would still have be written for the operating system. Not having to deal with x86 specific code and edian issues would simplify ports considerably.
goodjello has the idea
The idea of an OS X version that runs on x86 hardware is a good idea.
1) it promotes the development of OS X software.
2) it introduces the current windows user to the 'Mac' experience.
3) it would promote hardware purchases to get Native PPC speed.
4) it would provide a well-known alternative to the Linux and Windows experience on the x86.
Not risk free: 1) the item would compete with the Mac Mini but since the mini is an G4 machine it could be upgraded to G5 and handle the competition. 2) Since the x86 CPU architecture is not standard, a limited release through a known major vendor would be a sound move. If successful, other manufacturers would try to meet the architecture standard.
IMHO
The idea of an OS X version that runs on x86 hardware is a good idea.
1) it promotes the development of OS X software.
2) it introduces the current windows user to the 'Mac' experience.
3) it would promote hardware purchases to get Native PPC speed.
4) it would provide a well-known alternative to the Linux and Windows experience on the x86.
Not risk free: 1) the item would compete with the Mac Mini but since the mini is an G4 machine it could be upgraded to G5 and handle the competition. 2) Since the x86 CPU architecture is not standard, a limited release through a known major vendor would be a sound move. If successful, other manufacturers would try to meet the architecture standard.
IMHO