For all that we know, Jaguar is 64-bit clean, and Apple has simply been compiling it with -DPPC_32 rather than -DPPC_64 or -DPPC_FAT (I'm making up the names of #define constants here, but you get the idea).
OS X was built on a highly portable codebase. If Apple has been careful, most of the issues involved in preparing for a 64 bit CPU would have to do with extending and porting technologies brought over from Mac OS (QuickTime, Carbon, etc.), and they could have been done at the time, if a POWER4 machine was handy to use as a 64-bit PPC for testing purposes. At this point, it might really be as simple as throwing a switch at compile time. We don't know. But given the presence of CPUs which can function as 64 bit PPCs, given the IBM-stated fact that the 970 is derived from one of these CPUs (the POWER4), and given that most of the work in preparing OS X for 64-bit operation could have been folded into the work of putting it together in the first place, it's not at all beyond reason that the work is already done, and Apple has been waiting for the hardware to catch up.
Remember, the PowerPC ISA has been 64-bit since 1994. This is neither an abrupt nor an unforeseen nor a difficult transition. All IBM is doing is releasing a PPC that actually implements the complete ISA, rather than just the 32-bit subset.
Why does IBM all of the sudden support Alti-vec, is it because they've seen the progress of the G4 add-on and finally like it?
-walloo.
Some sites have reported IBM efforts to sell Apple some advanced CPUs for several years now with Apple being the one to say no. Apple have created their own problem in remaining with Motorola for so long with very little progress to show for it compared to other developments int eh industry.
Remember, the PowerPC ISA has been 64-bit since 1994. This is neither an abrupt nor an unforeseen nor a difficult transition. All IBM is doing is releasing a PPC that actually implements the complete ISA, rather than just the 32-bit subset.
And there have been 64-bit implementations since about 1997 with the PowerPC 620, POWER3, and POWER4. I would be astonished if Apple hasn't had at least a couple of these chips to play with at some point. They used to have a Cray supercomputer to play with, and that was completely incompatible with their production chips -- these PPC64 chips run the same architecture and its been obvious for years that eventually their descendents would be well suited for desktop use.
Some sites have reported IBM efforts to sell Apple some advanced CPUs for several years now with Apple being the one to say no. Apple have created their own problem in remaining with Motorola for so long with very little progress to show for it compared to other developments int eh industry.
I would take this with a grain of salt. The 970 project started about Y2K. Before that there were about 2 years where IBM didn't want anything to do with AIM or AltiVec and so weren't doing anything appropriate for Apple's use. This means Apple had a period of 2 years where they were saying no to a hypothetical IBM chip that didn't have AltiVec support, and access to Motorola's G4 which still looked strong at the time. But they were buying IBM G3s. By the time Y2K rolled around and the G4 was running into problems IBM was already getting the 970 project underway with AltiVec, and that was obviously intended for use by Apple. So its not clear to me when Apple was saying no to IBM about a chip that they could possibly use.
For all that we know, Jaguar is 64-bit clean, and Apple has simply been compiling it with -DPPC_32 rather than -DPPC_64 or -DPPC_FAT (I'm making up the names of #define constants here, but you get the idea).
It's an appealing thought, but remember Apple does release the kernel source under the name of Darwin, and expunging the 64-bit cleanliness would have been a major headache. Instead, they probably have a second kernel branch into which changes from the Jaguar (Darwin 6.x) and Panther (Darwin 7.x?) are merged.
It's an appealing thought, but remember Apple does release the kernel source under the name of Darwin, and expunging the 64-bit cleanliness would have been a major headache. Instead, they probably have a second kernel branch into which changes from the Jaguar (Darwin 6.x) and Panther (Darwin 7.x?) are merged.
But i would expect the 64 bit OSX to be of the 7Axx branch. Coinciding with Darwin 7 perhaps.
i wonder if the open source community is able to see any code-fragments which leed us to believe that 64bit-support is coming or already built in. any knowledge of this possibility in here?
There is a problem with browsing source code to find hints of the future.
You just don't know what is Plan A, and what is Plan B. As an example, DVD Player.app has a line disabling debug code on Pentium III, Pentium IV and Athlon CPUs. So Apple is obviously working (at least in a limited fashion) to keep Mac OS X portable to IA-32.
But does this mean Apple plans to move to IA-32? You can't tell from it.
Comments
Originally posted by Henriok
What I don't have is experience with porting operating sysems to other plattforms, who among us have?
Hmmm.... I wonder...
OS X was built on a highly portable codebase. If Apple has been careful, most of the issues involved in preparing for a 64 bit CPU would have to do with extending and porting technologies brought over from Mac OS (QuickTime, Carbon, etc.), and they could have been done at the time, if a POWER4 machine was handy to use as a 64-bit PPC for testing purposes. At this point, it might really be as simple as throwing a switch at compile time. We don't know. But given the presence of CPUs which can function as 64 bit PPCs, given the IBM-stated fact that the 970 is derived from one of these CPUs (the POWER4), and given that most of the work in preparing OS X for 64-bit operation could have been folded into the work of putting it together in the first place, it's not at all beyond reason that the work is already done, and Apple has been waiting for the hardware to catch up.
Remember, the PowerPC ISA has been 64-bit since 1994. This is neither an abrupt nor an unforeseen nor a difficult transition. All IBM is doing is releasing a PPC that actually implements the complete ISA, rather than just the 32-bit subset.
Originally posted by willywalloo
Why does IBM all of the sudden support Alti-vec, is it because they've seen the progress of the G4 add-on and finally like it?
-walloo.
Some sites have reported IBM efforts to sell Apple some advanced CPUs for several years now with Apple being the one to say no. Apple have created their own problem in remaining with Motorola for so long with very little progress to show for it compared to other developments int eh industry.
Originally posted by Amorph
Remember, the PowerPC ISA has been 64-bit since 1994. This is neither an abrupt nor an unforeseen nor a difficult transition. All IBM is doing is releasing a PPC that actually implements the complete ISA, rather than just the 32-bit subset.
And there have been 64-bit implementations since about 1997 with the PowerPC 620, POWER3, and POWER4. I would be astonished if Apple hasn't had at least a couple of these chips to play with at some point. They used to have a Cray supercomputer to play with, and that was completely incompatible with their production chips -- these PPC64 chips run the same architecture and its been obvious for years that eventually their descendents would be well suited for desktop use.
Originally posted by RBR
Some sites have reported IBM efforts to sell Apple some advanced CPUs for several years now with Apple being the one to say no. Apple have created their own problem in remaining with Motorola for so long with very little progress to show for it compared to other developments int eh industry.
I would take this with a grain of salt. The 970 project started about Y2K. Before that there were about 2 years where IBM didn't want anything to do with AIM or AltiVec and so weren't doing anything appropriate for Apple's use. This means Apple had a period of 2 years where they were saying no to a hypothetical IBM chip that didn't have AltiVec support, and access to Motorola's G4 which still looked strong at the time. But they were buying IBM G3s. By the time Y2K rolled around and the G4 was running into problems IBM was already getting the 970 project underway with AltiVec, and that was obviously intended for use by Apple. So its not clear to me when Apple was saying no to IBM about a chip that they could possibly use.
Originally posted by Amorph
For all that we know, Jaguar is 64-bit clean, and Apple has simply been compiling it with -DPPC_32 rather than -DPPC_64 or -DPPC_FAT (I'm making up the names of #define constants here, but you get the idea).
It's an appealing thought, but remember Apple does release the kernel source under the name of Darwin, and expunging the 64-bit cleanliness would have been a major headache. Instead, they probably have a second kernel branch into which changes from the Jaguar (Darwin 6.x) and Panther (Darwin 7.x?) are merged.
Originally posted by Anonymous Karma
It's an appealing thought, but remember Apple does release the kernel source under the name of Darwin, and expunging the 64-bit cleanliness would have been a major headache. Instead, they probably have a second kernel branch into which changes from the Jaguar (Darwin 6.x) and Panther (Darwin 7.x?) are merged.
Yeah, I did kind of forget about that. Heh.
Good point.
10.2.5 = 6Lxx
What ever happened to the branches 6Jxx and 6Kxx?
But i would expect the 64 bit OSX to be of the 7Axx branch. Coinciding with Darwin 7 perhaps.
Originally posted by Henriok
10.2.4 = 6Ixx
10.2.5 = 6Lxx
What ever happened to the branches 6Jxx and 6Kxx?
But i would expect the 64 bit OSX to be of the 7Axx branch. Coinciding with Darwin 7 perhaps.
i wonder if the open source community is able to see any code-fragments which leed us to believe that 64bit-support is coming or already built in. any knowledge of this possibility in here?
You just don't know what is Plan A, and what is Plan B. As an example, DVD Player.app has a line disabling debug code on Pentium III, Pentium IV and Athlon CPUs. So Apple is obviously working (at least in a limited fashion) to keep Mac OS X portable to IA-32.
But does this mean Apple plans to move to IA-32? You can't tell from it.
Barto