Apple has confirmed to the Register Panther, aka Mac OS X 10.3, will not be a fully 64-bit operating system. Greg Joswiak, Apple’s vice president of hardware product marketing, this week admitted that Panther will be a 32-bit operating system tweaked to support 64-bit addressing. Joswiak himself says: “The important thing for us [is] we didn’t want to create a separate OS that is 64 bits. What is essential is that this OS and this hardware will run 32-bit applications with no recompiling – it will just run them.“
Apple’s “First 64bit Desktop Platform” Won’t Run a 64bit OS
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
I’m not complaining. A Dual 2GHz Power Macintosh is still a Dual 2GHz Power Macintosh, 64bit OS or not.
MuHaha. 🙂
You should be complaining, with out a 64bit-OS the software is not going to take full advantage of the hardware.
if marketshare is as important as many believe, then the last thing Apple needs to do is get hung up on a 64 bit transistion, is “The Market” still not resolutely 32 bit?
Of course it’s not a 64 bit OS!!!
As has already been discussed ad nauseum on the various Mac sites, of course Panther isn’t pure 64 bit OS since it still has to run on 32 bit CPUs. This is nothing new. There’s going to be a period of transition.
http://macslash.org/article.pl?sid=03/07/28/1140212&mode=thread
huh? excuse me but the OS does take full advantage of the Proc. it services 32 bit apps in hardware, no emulation.
if you mean that it can not address 64 bit space…..um…so?
This has been discussed in many forums already, including this one and slashdot.
Everyone already knows this, but I still want to get my hands on a new G5….
Who cares if it is FULLY 64-bits or not. The article made it clear that programs will be able to do 64-bit addressing, unlike the previous articles on OSNews claimed. So if I want to run a huge program, it can natively address more than 4 GB. Apparently, there is a 32bit-64bit bridge built into the PowerPC 970.
“I’m not complaining. A Dual 2GHz Power Macintosh is still a Dual 2GHz Power Macintosh, 64bit OS or not.”
Not to mention the fact that the G5 is inherently faster G4 and P4… and in some tests, also the XEON.
The OS will be able to address 64-bits (or 40 or 48 or something huge), while individual apps can use an entire 32 bit space each. And I think newly compiled apps can even participate in the whole address space.
And it WILL include math libraries that allow 64 integer math, which is one of the few real performance increases that a 64 bit processor actually delivers.
So, tell me again, why would I care that it’s not a true 64-bit OS?
I’m the author of one of the previous OSNews articles on this subject.
What this new article confirms is that 10.2.7 and 10.3 will be the same in their support for large amounts of RAM.
What is still unclear is how these OSes will support “64-bit” addressing. Will they use programmer-aware hacks, as is necessary in 32-bit Windows to get an application to use more than 2/3GB of RAM, or will they transparently allow 64-bit compilation and execution of programs with only a compiler option, as on a true 64-bit OS? For example, just add the -m64 flag to gcc, as in
gcc -m64 largemem.c
From my point of view, once people start getting G5s, I can probably convince one of them to try to compile such a C code and see whether Mac OS 10.2.7 will run it or not.
The fact that Apple has not openly addressed this issue in its documentation leads me to believe that it probably will not work, which is unfortunate.
…is B.F.D.
Windows 95 had a bunch of 16-bit code in it and it took a few years to get that all worked out, but look at the improvement it was over 3.1.
At least Apple is improving their offerings. If they can’t get it all done in one day, who cares?
Good on you Apple.
Yeah we know. Do these analysts and journalists need to keep explaining this? Ofcourse it won’t, can you imagine Apple trying to manage a OS transition from 32bit to 64bit in the middle of all the other improvements and compatibility they are trying to ensure?
I think it’s better that for now we don’t see Apple converting Darwin into a 64bit OS we need other improvements first.
Ok, its 2003. I watched Beyond 2000 in the 90’s. Where are my flying cars and 64-bit Notepad.exe?
< This is a joke >
I got mac x 10.1
after 3 months Jaguar came out
Now panther 32bit
in 6 months the update to 64bit
BORIIINNGGGGGGGGGGGG and EXPENSIVEEEEEEEEEEEEE
There are two ways to do a bit transition. Some are better than others. Some examples are illustrative:
Solaris: When Solaris made the jump from 32-bit Super SPARC to 64-bit UltraSPARC (which is very similar to the G4 -> G5 jump in that the UltraSPARC can run 32-bit code natively) it recompiled the OS itself in 64-bit mode, and provided a modified program loader, a parallel set of system calls, and a set of libraries compiled in 32-bit mode to support older 32-bit apps. This is the best model. New apps get full support for a 64-bit address space while old apps stay supported. This is also relatively easy to do, but requires that the OS is 64-bit clean to begin with.
Windows: When Windows 9x made the 16-bit protected mode (Win16) to 32-bit protected mode (Win32) jump, large parts of the core OS stayed 16-bit, while a 32-bit API was provided on top. This meant that the OS code was littered with functions called “thunks” which adjusted parameters and such to allow 32-bit code to call 16-bit functions. This was, to say the least, buggy and slow.
Now, I don’t think Panther is going to adopt either of these models wholesale. In fact, if Panther itself is 32-bit, the most likely scenario is that it won’t even attempt a decent 64-bit implementation, waiting instead for a full 64-bit transition in a later version of the OS. What will most likely happen is that Panther will be > 32-bits in the same way WinXP on x86 is greater than 32-bits — it will allow the OS as a whole to use > 4GB of physical RAM, and will allow apps to create less-than 4GB memory windows that can be mapped onto greater-than 4GB chunks of memory.
If Panther itself remains compiled in 32-bit mode, it would be difficult for it run applications compiled in 64-bit mode. Consider this: A 64-bit application issues a read() system call to read some data from a file into a memory buffer. If the buffer resides in a portion of memory above the 4GB barrier, it is extremely difficult for the 32-bit code inside the OS to put data in that buffer. If Panther did allow running apps compiled for 64-bit mode, it would have to put an additional requirement that no pointers to memory above 4GB could be passed to library functions or system calls. The apps could treat these memory areas merely as buffers for its own uses.
As has already been discussed ad nauseum on the various Mac sites, of course Panther isn’t pure 64 bit OS since it still has to run on 32 bit CPUs.
Solaris still runs on systems with 32-bit CPUs, yet it is a full 64-bit OS.
The solution is to provide 32-bit and 64-bit libraries as well as a 32-bit and 64-bit kernel, then choose which kernel to use based on the platform.
With G5s shipping with hard drives with a minimum size of 80GB, bundling a hundred megs of extra libraries doesn’t really seem like much of a concern.
This will most likely happen some time in the future. Apple will need to design and modify a number of programs and tools before this can happen. The first step would be designing a Mach-O ABI for 64-bit PPC. This ABI would then need to be implemented in gcc, after which time a 64-bit build of the kernel could be made. After this, all the dev tools would need to be modified to support the 64-bit ABI. Applications built with the 64-bit ABI could also fall back on the 32-bit ABI thanks to Mach-O’s fat binary support.
I don’t think Apple taking their time is bad, provided that when they do (hopefully) release a full 64-bit OS it is done right.
Mac OS X just keeps getting better and better 😛
How hard is it to have two different CDs one with a version compiled for 32bit addressing, and one for 64bit addressing. Granted there would be some code changes, but Linux, Solaris and heck even Windows has been distributed for multiple CPU architectures. This seems like apple is slacking off, and I wont be buying a G5 till I can get 100% performance out of it.
They are making it a smooth transition, next year, with X.4, That should end up being the true 64-bit OS on 64-bit hardware.
And we all saw how well that worked!
and you would have seen from the beginning that Apple was shipping the G5 was a 32 bit Mac OS X that would support programs using 64 bit registers. It was ALL OVER the place.
They are doing this the RIGHT way as compared to Intel and AMD. Neither of these will be able to run 32bit programs. So most of what they will be used for are SERVERS. Since Apple’s server sales is almost invisible compared to their desktop sales this wouldn’t make sense. Especially when they have such a HUGE need for faster desktops.
Since desktops are the focus, they need to make sure 32 bit apps will run. There is no easier way than to have a 32 bit OS to run 32 bit apps. (Notice I said easier).
As Apple has time and more and more people buy G5 Macs they will shift more and more parts of Mac OS X to 64 bit. Then one day Apple and maybe IBM will come out with a fully 64 chip and OS and everything will be 64 bit.
Look at the past. It tells us about the future.
“How hard is it to have two different CDs one with a version compiled for 32bit addressing, and one for 64bit addressing.”
It is not difficult, just very time consuming as *all* the API’s that potentially need to be extended to 64-bits need to be updated with a parallel set of system call in the kernel along with parallel user land runtime libraries. It takes time, but at least the initial version of Panther will support the ability for applications that need 64-bit addressing to be able to utilize it.
Basically it will take a lot of engineering effort to properly translate all the API’s, get them solidified and documented, get the code built and more importantly get it all tested to ensure that existing 32-bit apps still function correctly and new 64-bit apps also function correctly.
Heck, even 64-bit Linux didn’t happen overnight, and it took quite a few years for the ports to become stable.
The way I see it, Apple is approaching this the right way.
Is OS News a tech site or is it only “anti-Apple”?
What’s the difference between a “not full 64 bit OS but that supports 64 bit addressing” and a full 64 bit OS? Is it a full 64 bit OS only if it uses 64 bit PID??
To me it’s ok if it supports 64 bit addressing and uses 32 bit PIDs 🙂
(obviously 32 bit applications will continue to see 32 bits addresses, that’s “by design”)
“They are doing this the RIGHT way as compared to Intel and AMD. Neither of these will be able to run 32bit programs. So most of what they will be used for are SERVERS.”
Where did you get that from? Intel’s Itanium processors have an emulation layer. Sure, it is not the fastest, but it does run 32bit programs.
AMD’s Opteron processor can run 32bit and 64bit programs natively. And so will the Athlon 64 when it comes out.
Intel might even implement AMD64 in their processors. So both camps can definitely run 32bit programs.
the marketing hype from appl promised a 64 bit system; which is technically misleading. that’s why the response, and importance of this revelation.
They are making it a smooth transition, next year, with X.4, That should end up being the true 64-bit OS on 64-bit hardware.
and what mainstream app would you run with a 64-bit OS (on a Mac) in the immediate future? software makers need the time to make the transition too, and make their products available for purchase..
it’d be awfully swell to have a fully 64-bit Panther running on a brand new G5 the second they start shipping, but aside from beautifully-supported hardware, you wouldn’t be able to do much with it..
When the specs on the G5 came out no one expected for it to be real. So a few days afterwards Apple announces a 64-bit desktop system. No mention was made of the OS being 64-bit but some of the Macheads assumed that the 64-bit hardware would mean it would be shipping with a full 64-bit desktop which is not the case.
It seems that EVERYONE seems to have high expectations for Apple even the haters. To the glee of the haters, Apple will NOT be shipping 10.3 as a 64-bit OS.
I do agree with the consensus so far that Apple is going about transitioning to 64-bit the right way providing some of the advatanges of a 64-bit OS and without sacrificing speed and compatibility. In Apple’s implementation there seems to be more advantages than disadvantages.
Agreed OS X keeps getting better and better, period!
Someone posted this on MacNN. Another unsupported comment, but it goes in the lines of earlier posts that Mac applications will NOT be able to use more than 4GB of RAM, or in practice more than 2.25GB or so.
10.2.7 and Panther use 32b effective memory addressing. This represents the number of bits that can be used by Applications to address memory. In other words the normal Application address space is 4GB in size. Panther (and 10.2.7) do have an enhanced virtual memory system that at least allows 8GBs of physical RAM to be addressed. So the VM system can back its virtual memory space with at least 8GB of physical RAM (in theory it supports much much larger then this… i think 48b addressees don’t have my notes handy). At the moment I have not heard if and how one could special case an Application to support a larger effective address (it most likely will have to wait until post Panther).
http://www.digitalvideoediting.com/2003/06_jun/features/cw_macg5_in…
🙂
Reminds me of Microsoft’s 32-bit file support and their win16 to win32 transitiion. Good for business, but not very technically challenging.
But I’d rather run Linux anyway, so go fig.
“” rel=”nofollow”>http://www.digitalvideoediting.com/2003/06_jun/features/cw_macg5_in…
Anyone that reads the following link should consider the fact that it is written by Charlie White. Anyone that has followed Mr. White’s Mac vs. PC comparisons in the past, is undoubtedly aware of his affection for PCs and the fact that he loathes the Macintosh and works hard to discredit Apple’s product offerings at every corner.
For example, last July, only days before Mac World New York, Mr. White did a comparison between an end-of-life model PowerMac and a cutting edge PC, a PC, which had coincidentally been introduced only a week or two before his article was published. As a matter of fact, one of the PC models he included in his report was one that had not even started shipping yet!
Surely Mr. White knew that new PowerMacs were coming, but he chose to do his comparison with 6-month-old Macs and brand new PCs.
Worth noting is the fact that one of the PCs that he chose to use was a new $4000+ custom dual Athlon Digital Video system; yet somehow he neglected to mention the significant price difference in his comparison.
In November, Mr. White did another comparison. Why did he wait until November to update his benchmarks? Apparently he was waiting until he could get his hands on a 3Ghz Pentium 4 Dell system, which of course was conveniently released at that time.
Either through error or intention, he misrepresented the pricing on both the Dell and the Mac. Although he quickly corrected the Mac price, the Dell price is STILL shown as being under $3000 even though the tested system had a price over $3200.
We are now nearing the end of July and there has been no update to the benchmarks within Charlie’s reports, this despite the fact that new PowerMacs and PowerBooks were introduced almost 6 months ago. I can only assume that he must be waiting for something new to come from Intel or AMD.
AMD’s Athlon-64 processors have been repeatedly postponed and the only thing new from Intel is the company’s Centrino chip. Knowing this, you would think that it would be appropriate to create a dual for video editing professionals on the go by comparing Apple’s latest PowerBooks to a PC notebook utilizing Intel’s new Centrino chip.
Knowing Charlie’s biases, I doubt we will ever see such a comparison, simply because the PowerBook would win hands down… this for both performance AND price. Charlie would never consider publishing benchmarks where a Mac came out on top.
He changed the benchmarks used between his July and November articles? Why? Perhaps the reason is because the new Mac leap frogged the previous Mac and PC benchmarks.
and playing with the G5 the Apple fellow came up to “help me”. I said how fast it was and how pokey my machine was going to feel when i got home. He started to make a comment about “yes, and this OS isn’t even really optimized for the G5 yet”. Stupidly it didnt click till later that i should have asked what he meant since he was just full of factoids. I, on the other hand, was so mesmorized with launching every app possible (at the same time) and trying to find something that actually required the dock icon to bounce a full bounce, that i could not be bothered to find out what he was talking about. oops. I’ll chalk that one up to my crappy people skills and just how darn fast that machine felt compared to my G4 400mghz at home.
Biased or not, the point is made – if an IBM representative cannot give a convincing response to the question on whether Apple G5 is the first 64-bit desktop computer, and Apple’s touting 64-bit as the main advancement against the PC counterparts, isn’t that claim a hype? Of course, we all know that, don’t we?
“Biased or not, the point is made – if an IBM representative cannot give a convincing response to the question on whether Apple G5 is the first 64-bit desktop computer, and Apple’s touting 64-bit as the main advancement against the PC counterparts, isn’t that claim a hype? Of course, we all know that, don’t we?”
Well, yes, and no. If it is so clearly hype, then why are you pissed off? I don’t see many people ranting about BuyMusic despite the fact that l3ess than 5% of tracks are $.79… Okay, maybe you defend that because they say “From” $.79.
But what about in the ToCs where it says despite saying buy, purchase, etc.. you do not own any content, hmm?
Anyway, my point is: if it is just hype, then how can you say that you are being deceived?
If it isn’t just hype, then it’s very precise marketing language. In which case White is just an @ss because despite his claims, Boxx calss their system a workstation, AMD doesn’t consider their chip a desktop proc, etc…
As for Apple, they say they have the fastest 64-bit desktop system. I think that’s valid. It’s a 64 bit chip. The OS will handle 64 bit addressing — the main and most sought after advantage of 64 bit-ness. And when anyone says anything about “speed” and computers–you are aware that there are a number of taks and areas to be speedy at, right?
So where’s the confusion? Why is Apple to blaim because you’re an ignoramus?
As to your own point, have you even read your own propaganda? What Rubenstein and the dude from IBM don’t know about is the status of a gray-box assembler’s naming of a product. Why the hell should IBM reps need to know what AMD and some tiny company are marketing their products as? How does that prove anything?
“Biased or not, the point is made – if an IBM representative cannot give a convincing response to the question on whether Apple G5 is the first 64-bit desktop computer”
You’re forgetting that the question was loaded to begin with. It would be akin to someone asking George Bush if he thinks men or women are better leaders. Any response he gives can be used against him so I would imagine that we would give a half assed remark until it the question could be more properly addressed. IBM was in the same boat when the question was posed to it.
The fact that Charlie posed the question at all showed his bias even before he write his article.
“and Apple’s touting 64-bit as the main advancement against the PC counterparts”
Apple is touting speed… something that the G5 will definately bring. If they should happen to benefit from the fact that the average consumer doesn’t know that speed comes from the chip technology itself and not its “64 bitness” then all the better. Keep in mind that Apple lost a lot of mind/market share as a result of intel purposefully playing upon the public’s ignorance about MHz/GHz and speed.
The argument about asking Bush’s opinion about value of the men and wonem is irrelevant here. Opteron is already available on the market while G5 is not and this is the fact not the opinion.
Now opinion is what is here workstation and what is desktop. In case of G5/panther is difficult to say as the same hardware is used in both cases. The size of HD or RAM does not matter, but installing ECC and high-end SCSI plus 3dlabs Wildcat or nVidia Quadro4 would make difference. Neither ECC nor high-end video cards are available for OS X. So what would make OS X workstation? On the other hand my “desktop” could be dual Opteron 244 with 64-bit operating system installed (SuSe or WinXP) plus ofice/browser/video editing/whatever apps. Adding IDE HDD plus ATI 9800PRO/nVidia 5900 and DDR (non-ECC) RAM would make standard desktop.
NO ONE who comes here should be listening to this marketing BS.
Oh, and in apple’s defense they have to use the 64bit BS to counter the into mhz BS.
It achieves equilibrium between the two companies’ marketing garbage.
change “into” to “intel”
MP: “On the other hand my “desktop” could be dual Opteron 244 with 64-bit operating system installed (SuSe or WinXP) plus ofice/browser/video editing/whatever apps”
But even you are “quoting” your “desktop” and you would have to build it. How does that discredit Apple? Apple has said they are the only manufacturer shipping a 64 bit desktop system. There are truly no manufacturers doing so (particularly tier-1). So where’s the bull?
Well, yes, and no. If it is so clearly hype, then why are you pissed off?
Did I sound pissed off? I didn’t even say Apple is worse than any other company or whatever.
if it is just hype, then how can you say that you are being deceived?
Did I say I was deceived?
How does that prove anything?
It proves that they know little beyond their own product. It proves that they are disconnected from their marketing department, even though they are doing some marketing in that interview.
Apple is touting speed… something that the G5 will definately bring.
Ok, I was wrong. It’s not their main marketing line. But still, they do say “The Power Mac G5 is the world’s fastest personal computer and the first with a 64-bit processor“.
Now opinion is what is here workstation and what is desktop.
That is true. Apple’s offering start at 2,000 (or 1,800 after removing this and that). BOXX’s starts at 2,500, but comes with Quadro4. Price makes a workstation a workstation? Maybe…
How does that discredit Apple? Apple has said they are the only manufacturer shipping a 64 bit desktop system. There are truly no manufacturers doing so (particularly tier-1). So where’s the bull?
Huh? Umm… ok.
Anyway, my lack of elaboration will definitely create too much misunderstanding. I started out with “Apple lied about being the first 64-bit desktop computer” which after some critiques here should be changed to “Apple didn’t know it wasn’t the first company to offer 64-bit desktop computer but still not acting to resolve this weird situation they put themselves in”.
Oh well, I knew I shouldn’t have started this anyway. If you are offended by my posts, and think I’m ignorant, just do what you should do with trolls – ignore me.
“Anyway, my lack of elaboration will definitely create too much misunderstanding. I started out with “Apple lied about being the first 64-bit desktop computer” which after some critiques here should be changed to “Apple didn’t know it wasn’t the first company to offer 64-bit desktop computer but still not acting to resolve this weird situation they put themselves in”.”
No, this is completely wrong. I agree that haggling over the difference between a desktop and workstation is weak, but the most basic concept would be: accept the manufacturer’s definition of itself. Boxx refers to their product as a workstation. AMD refers to the chip as a workstation/server chip. So, if Boxx and AMD don’t think they are the first to produce a 64 bit desktop, who is White to say that Apple is wrong. Why do you agree with that idiocy.
“Huh? Umm… ok.” Don’t know what that means–are you suggesting that I’m wrong? No one to this day has yet pointed out a complete boxed system w/OS that is advertised and sold by the manufacturer as a 64 bit desktop. Please prove me wrong.
“I started out with “Apple lied about being the first 64-bit desktop computer” which after some critiques here should be changed to “Apple didn’t know it wasn’t the first company to offer 64-bit desktop computer but still not acting to resolve this weird situation they put themselves in”.”
All other points asside, they ARE selling the first 64 bit desktop computer… which makes all other points moot anyways.
The BOXX system is not a desktop computer. They way this is determined is by the OS that runs it. OS X runs on the G5 and the iMac/eMac. The BOXX system runs WinXP professional. Consumer hardware would be running XP.
IMHO, it is a silly discussion.
Can you see that?
Before everything, as it is a very partisan discussion… I use PCs. I used Macs for some time in Europe. Will not be using again in some time, because I do numerical calculations and I live in third world. So I love AMD!
But…
I think Apple is saying no lie. It has now a de facto 64 bits enabled desktop machine. It can be a genuine desktop machine.
I can also buy an Opteron and install 64bits Linux on it and do heavy numerics on it. It will not be a desktop though. It can be in a near future. It is not next month!
What I think is important is:
(For Mac users):Is it going to be a FASTER desktop??? Probably yes.
(For the majority of human beings only): Will it be affordable??? Probably no! Specially outside US and Europe, it will be a huge white-elephant that nobody will need and almost nobody will be able to pay.
If you let me say something about Apple, from my thinking:
When it comes the news that Apple reconsidered its elitistic approach and think about bringing real competition for the masses, therefore really doing something different, thinking REALLY different, I will applause.
We in third world are very happy that AMD (of course, a good side-effect) released us from the Intel monopoly. And we buy lots of its chips (remember some years ago it was a tiny player).
I will be very happy when Apple thinks different, like that.
It would earn a lot of money and get market share (as AMD did). I want Apple with us. Unfortunatelly… It is not.
“It proves that they know little beyond their own product. It proves that they are disconnected from their marketing department, even though they are doing some marketing in that interview.”
IBM’s suppos’d ta know ’bout BOXX? Boxx? For freakin’ real? IBM?
Are you suggesting that it should be standard practice to interview a company rep, ask them questions about some other obscure little companies products, and if they don’t know about it, then they are a poor company with a poor strategy and poor products? This is retarded. Unfortunately, you are so trapped in your own bias that you don’t understand how retarded that is. I wish I could interview Ballmer, or you for that matter, and ask you obscure questions about Apple software… See what kind of answers you know of.
“The BOXX system is not a desktop computer. They way this is determined is by the OS that runs it. OS X runs on the G5 and the iMac/eMac. The BOXX system runs WinXP professional. Consumer hardware would be running XP.”
Rubbish. XP Home does not support dual processors so XP Pro must be used on any dual processor system. XP Pro supports dynamic volumes, mulltiprocessors and IIS services. The user experience is on Pro and Home is identical. I have XP Pro on my (very non workstation) 1GHz Athlon.
The G5 is not a workstation classs machine because it doesn’t support ECC RAM or a professional graphics card.
Dual 2GHz, 64bit, RISC cpu’s = ownij.
who gives a fuck if the OS isnt pure 64bit right out of the bag…
If the new PowerMacs cain PC’s with an optimised 32bit OS now- a few more Revs and we will all see Mac’s taking the speed crown- once again.
And please dont feed me that shit about “64bit doesnt give any speed increases”.
I forgot- PowerMacs have already taken the speed crown once again :]
But the problem with the question of a 64 bit os or not, is that what do you call a 64 bits os, and i would like to ask this question to osnews?
We know now that Panther support the 64 bits adressing and that some mathematical libraries has been written to support the 64 bit intger unit of the G5 (ok a few people will need this!!!).
And in the article of theRegister, they say that panther may allow applications (or process) to adress more than 4 GB or ram and by the way in the last financial results conference, a journalist asked Apple if panther support the 64 bits applications, and the answer was ……yes! Where have you been guys?
So if Panther support the 64 applications, what is the problem?, what else do uou need to have a 64 bits os. If panther does so, we can use the main advantage of the 64 bits computing, which is to access to more data, and memory.
So i ask to osnews, why this provocative statement: Apple’s “First 64bit Desktop Platform” Won’t Run a 64bit OS.
Because first you seem that you don’t know what a 64 bits os is, and on the other hand you make a statement that relies on rumors, ……..
.
Okay, I’m only going to say this once more.
64-bit have NO advantage to existing Mac users. Why? Let’s see what 64-bit is.
64-bit means there is 64 different combinations of input and one output in all gates except for “NOT”. What does it bring to computing? Those computing with excessively large amounts of information would see the benefit, as they require the amount of inputs. However for normal desktop computing, it is totally unneeded, and 32-bit works fine.
As for Panther not being 64-bit, it is atually good. You don’t have to be 64-bit in order to optimize for 970. There is absolutely no technical reason for that, because (1) Mac doesn’t need all the extra flexiblity 64-bit offers, just a few and (2) Panther may actually be slower if they become 64-bit for no technical reason.
Now, this can’t be compared with the 68k to PPC transition where Mac OS was super slow because it was mainly in 16-bit 68k code. PPC didn’t natively support 68k code, therefore emulation was needed. Because of this emulation, it was slow. And also the fact that during the move, Mac needs exceeds the limitations of 16-bit, not similar to this case.
In other words, there is no advantage having Panther is pure 64-bit. Absolutely none. The disadvantage is that Apple would spend more time getting Panther out of the door, more money on QA for absolutely miniscule, if any performance gain.
“PPC didn’t natively support 68k code, therefore emulation was needed. Because of this emulation, it was slow.”
Yes, PPC emulated 68k code, but because the new PPC chip was so much faster, it actually ran the emulated 68k code FASTER than what it was able to do nativly. You are wrong about the slow speed
The BOXX system is not a desktop computer. They way this is determined is by the OS that runs it. OS X runs on the G5 and the iMac/eMac. The BOXX system runs WinXP professional. Consumer hardware would be running XP.
I’m getting extremely pissed off with this argument. Now, can anyone tell me in any way that consumer features within XP Home can’t be found on XP Pro? I’m using XP Home on this Duron machine, on my laptop I have XP Pro, and although I can name dozens of features XP Pro has over Home, I can’t name one Home has over Pro.
Then there is ease of use: they are identical in UI. The control panel is the same. The default options are the same. XP Pro has some extra control options from Windows 2000, but it is well hidden. In other words, the consumer can use as productively XP Pro as he can for Xp Home.
So what’s XP Home? It is a striped down version of XP Pro. It is marketed towards people that don’t need features that is available on XP Pro. Now, as for BOXX, Windows XP Pro is used because it requires a feature XP Home doesn’t have – dual processor support.
And the last I checked, consumers aren’t really gung ho about workstation features, mainly dual-processor and 64-bit. They don’t need it. And such machines falls way out of range with the average consumer machine brought at $800. And the last I checked, consumer apps (even iApps) don’t need dual-processor, and the difference of performance between workstations and normal lower end machines is extremely minimal.
And G5 isn’t marketed towards the home market. OS X has two versions, one OS X, the other OS X Server. Apple has absolutely no need to have two versions of OS X for both consumers and workstations as it doesn’t make its money from OS sales, rather machine sales. OS X has both consumer and workstation features in it. In other words, it is only comparable with XP Pro, not XP Home. G5, in its ads, in its pricing, in its benchmarks just show one thing: they are targeting the workstation market, not the consumer market.
Yes, PPC emulated 68k code, but because the new PPC chip was so much faster, it actually ran the emulated 68k code FASTER than what it was able to do nativly. You are wrong about the slow speed
Not the way many people remembered it. I can’t say for sure, I wasn’t a Mac user then (or now). But everyone I know who bought a early PPC machine from Apple reported it was slower until later versions of Mac OS (namely 8.5, IIRC).
“Now, as for BOXX, Windows XP Pro is used because it requires a feature XP Home doesn’t have – dual processor support.”
So, it requires XP Pro (not XP consumer), BOXX refers to its computer as a workstation, and AMD refers to its chip as a work station chip. The fact that some may use this work station as a desktop computer doesn’t change the fact that its still a workstation. More to the point…. it doesn’t make Apple’s statement that they shipped the first 64bit personal computer a lie.
“Such machines falls way out of range with the average consumer machine brought at $800.”
Actually, thats the low side of average. I recently read that the average consumer purchase was between $799 and $1,800. Coincidentally, the high price point is the dollar amount that Apple sells its G5 at. So, yes, the G5 is a desktop computer, and yes, its priced right for the consumer.
“And the last I checked, consumer apps (even iApps) don’t need dual-processor”
And yet they are certinly benefited by them.
the difference of performance between workstations and normal lower end machines is extremely minimal.”
What? Are you saying that the performance difference between a eMac (for example) is minimal in comparison to a XEON or AMD, or are you trying to say that the performace difference is negligable because the average consumer supposedly only writes email, browses the web and writes letters?
Either way, you’re wrong on both counts.
The speed difference is there and is significant. Secondly, the average consumer is doing more than writing letters, browsing the web and checking e-mail.
“OS X has two versions, one OS X, the other OS X Server.”
OS X server isn’t a workstation OS (like XP Pro) its a server OS.
OS X, is a consumer OS and a workstation OS.
“OS X has both consumer and workstation features in it. In other words, it is only comparable with XP Pro, not XP Home.”
Its only comparable as far as its workstation capeabilities. Because XP Pro is the workstation OS and not the consumer OS, its not the same. Microsoft doesn’t have a product to match OS X.
“G5, in its ads, in its pricing, in its benchmarks just show one thing: they are targeting the workstation market, not the consumer market.”
The low end G5 is priced for the high-end consumer.
It’s specs show that it can meet the needs of the most demanding consumer as well as the most demanding professional.
So far, the only ad for the G5 is that of a CONSUMER using the G5 at HOME.
I don’t see what the problem is. There’s not a lot of 64bit support out there yet.
What do you think people will be running on the Athlon64 when it comes out?? I’m guessing 32bit Windows or Linux!
So I don’t see this as a problem. I’m sure Apple will come out with a pure 64bit version of OS X eventually… I bet they’re working on it right now!
So I see this as a total non issue.
“Not the way many people remembered it. I can’t say for sure, I wasn’t a Mac user then (or now).”
You almost have a point if we were to consider the short-lived 601 PPC chip. The 601 emulated the code at the same speed of previous 68k processor. Every PPC chip thereafter ran 68k code faster.
“But everyone I know who bought a early PPC machine from Apple reported it was slower until later versions of Mac OS (namely 8.5, IIRC).”
I think you may be misinterpreting what they said. The 68k code ran slower than PPC, but not slower than native 68k.
“64-bit means there is 64 different combinations of input and one output in all gates except for “NOT”. What does it bring to computing? Those computing with excessively large amounts of information would see the benefit, as they require the amount of inputs. However for normal desktop computing, it is totally unneeded, and 32-bit works fine. ”
What the heck does the first statement mean? 64-bit means in the high level sense the processor can address 2^64 bits of memory. Which means all pointers and possibly registers are 64-bit (they probably have some quad precision floating point registers, I don’t know the ppc970).
Now you can address only say 48 bits of physical memory and still export a 64 bit adress space to the processes but map them to the 48 bit physical addresses in the mmu.
The kernel can reside in say a 32 bit privileged address space and still allow for 64-bit address spaces for the individual processes. So that means when the krenel executes all translations for the kernels Virtual Adresses -> Physical Addresses would be 32-bits. When a context switch happens to a process with the 64 bit Address Space the kernel loads the right translations in the mmu and the process can address > 4GB.
So you have a OS which is not fully 64-bit but can support 64-bit Adress Spaces for userland processes. There should be no drastic performance issues but only those of supporting maybe two different page tables and may be some overhead during context switching.
I wish people would understand the issues involved in making a transition from 32-bit to 64-bit is a a major change in the OS and a gradual change is more benifitial to customers. As long as the userland processes can address > 4GB why this debate.
Has OSNEWS, EVER posted a postive story about Apple?
Of course not.
Every Apple story posted here has a somewhat pessimistic tone to it.
Actually, thats the low side of average. I recently read that the average consumer purchase was between $799 and $1,800. Coincidentally, the high price point is the dollar amount that Apple sells its G5 at. So, yes, the G5 is a desktop computer, and yes, its priced right for the consumer.
I don’t know where you get that statistics, but the last I checked (a couple of months ago), about 30% (rounded up 🙂 of PCs sold worldwide are between the range of $700-800. And there are far more PC sold between the $500-600 and $600-700 range than those sold above $1000, especially in Asia.
I don’t have the stastitics off-hand, and question the need to search for it.
And besides, the highest end smart phone is around the same pricing range of an low end PC. Would that make the smartphone equalivent to a low-end PC? Hardly.
And yet they are certinly benefited by them.
Not by a whole deal. The main performance difference, the one that would be appreciated, is the one amongst professionals. Sure, I can save a second or more in loading Word, or 1/2 a minute in rendering this movie file – but those are extremely negliable if you are not doing it at a larger scale.
What? Are you saying that the performance difference between a eMac (for example) is minimal in comparison to a XEON or AMD, or are you trying to say that the performace difference is negligable because the average consumer supposedly only writes email, browses the web and writes letters?
As well as manage his digital photos, watch a DVD or two, have a webcam conversation, etc. A friend of mine that just migrated to New Zealand is having webcam conversations with his family here in Malaysia – on his 3 year old K6 machine. He doesn’t complain about speed.
And the last I went into an Apple store for a demostration of iMovie, it was on a eMac. It was very fast, I doubt I would appreciate the speed difference brought by G5 unless I was doing more complicated stuff and using higher-end software like FCP.
The speed difference is there and is significant. Secondly, the average consumer is doing more than writing letters, browsing the web and checking e-mail.
Quote me where I said the average consumer is doing more than writing letters, browsing the web and checking e-mail. In fact in other threads, you can see me saying the exact opposite.
OS X server isn’t a workstation OS (like XP Pro) its a server OS.
Your comprehention skills must be bad, cause I never said it was.
OS X, is a consumer OS and a workstation OS.
Precisely my point. And you saying that just because the PowerMac comes with normal standard OS X it is a consumer machine is also wrong. How would OS X be a workstation OS if the very top end machines are considered as consumer ones?
Its only comparable as far as its workstation capeabilities. Because XP Pro is the workstation OS and not the consumer OS, its not the same. Microsoft doesn’t have a product to match OS X.
[man, i’m getting tired of this argument] Actually, it does. It is called XP Pro. Just like OS X, it has consumer features and workstation features. It can fit any market segment OS X can. I’m not talking about the nitty gritty of features and niche, rather target market. XP Pro, while marketed for the workstation and enterprise market (the only markets that would appreciate the features), it can fit any market where OS X could.
In other words, if I want to give my laptop to a person that is in the consumer class and does consumer things, I don’t have to reformat and install Windows XP Home for him to be productive. He would be as much at home with XP Pro as he would with XP Home. Now, if he would to buy a new machine, he would probably go for Home. Why? Because it is cheaper, and he doesn’t need Pro features.
The low end G5 is priced for the high-end consumer.
There is no such thing as a high-end consumer. Consumer, as its popular defination here on OSNews (I’ve been around for a long time) regards to the lower segment of the market. People that don’t mainly use machines for work or hobbyist purposes.
That person can be a gamer or a excessive IM user. It doesn’t matter.
You probably meant the hobbyist market (i.e. extreme gamers, open source developers, etc). These people are not nitwits on computers. They know more or less about computers as the average OSnews geek. They don’t answer “huh?” when asked about Linux, freeBSD, etc. They understand what are their altenatives, rather than going for the mainstream products.
It’s specs show that it can meet the needs of the most demanding consumer as well as the most demanding professional.
The most demanding 3D professional would probably ask for specs like these:
Integrated vertex processing engine
Integrated image and texture engine
12-bit per component color and alpha, (Octane2 V6, V8, V10, V12) double-buffered (Octane2 V8, V12) 16-bit 2 buffer (Octane2 V12 only)
24-bit eye space Z buffer and 8-bit stencil
10-bit digital to analog (DAC) display interface
Multiple concurrent visuals (8-bit window ID)
V6 and V10: 32MB graphics memory, including up to 8MB texture memory
V8 and V12: 128MB graphics memory, including up to 104MB texture memory
And PowerMac G5s don’t even come with QuattroPro out of the factory. They are catering towards the low-end workstation market. A market where the demand isn’t too high (like those served by Sun and SGI), neither are their consumers as they use their machines purely for work. These people can be from low-end video producers to high-end 2D graphics designers to high-end sound technicians.
And why are PowerMacs priced so low? It is because PC workstations are priced slightly higher. This is always the case when Apple release a new product altogether (e.g. iMac when first released, in comparison Dells and gateways of the same price). Give it a few months, and you would see PCs become more cheaper (especially when Prescott is released).
Now, why did Apple use “the first 64-bit *desktop*”? Because it can’t claim being the first 64-bit workstation (that would be an outright lie). But if PowerMacs can be considered as desktops (as in lower end machines), then PC workstations can be seen in the same light. Especially so since most of them are 32-bit, support less than 4gb of RAM, etc.
So what’s you defination of desktops and workstations?
I rather not debate about this since both of us would be fighting it out indefinately. Now, even though if you have a point that emulated 68k was faster than native, it was still slower than PPC. Now, I don’t see proof that 32-bit PPC apps optimized for 970 would be any slower than one in 64-bit. And that’s my point – there is no emulation, the app (or OS in this case) doesn’t need it, so why the fuss?
> I wish people would understand the issues involved in making a
> transition from 32-bit to 64-bit is a a major change in the OS and a
>gradual change is more benifitial to customers. As long as the
> userland processes can address > 4GB why this debate.
It’s far from obvious that userland processes in Mac OS 10.2.7 and 10.3 can address more than 4GB in a practical manner. Certainly no online Apple documentation discusses the issue. For me, that’s what I am waiting to hear. I am afraid that the answer is “no”, but we will see.
Many times. Check out the news on the aftermath of WWDC. Or the ones about iPods, Apple Music Store, etc. Or even just recently on Safari market share increasing.
“It’s far from obvious that userland processes in Mac OS 10.2.7 and 10.3 can address more than 4GB in a practical manner. Certainly no online Apple documentation discusses the issue. For me, that’s what I am waiting to hear. I am afraid that the answer is “no”, but we will see.”
Did my post mention MacOS X? How did you draw your conclusion. I was merely stating that a OS need not be fully 64-bit. As long as it can support userland apps with 64-bit address spaces from the users perpective it is a 64-bit OS. Everything else is implementation details. The OS’s job is to hide those details from the user.
I am not sure if the OS shipping with the G5 will implement such mechanisms but there are subtle differences between a fully 64-bit OS an OS in which apps can address 64-bit. The differences however are more important to the engineering team designing future hardware. The transition should if properly implemented be tranparent to endusers. We will se how Apple does things.
I just found this one and it’s probaly a repost but… ITS SO FUNNY
http://www.thehoucks.com/happynowhere/Apple_Switch_Parody_DivX.avi
Yes there are other things than just memory addressing for 64bit stuff in the kernel. I admit that the kernel does not have to be a 64bit O/S to support these things, but I think it does indicate how far away MacOSX is from being a true 64bit O/S.
Other than memory, the other obvious thing that needs 64bit addressing is a filesystem. Things like filesystem and file sizes clearly need to be 64bit, offsets into a file as well. But there are other things that MacOSX has kept 32bit
that it exports to userland, like an ino_t.
Transitioning this to 64bit is not easy will require 64bit versions of functions like stat that pass back 64bit values in the stat structure.
Have Apple said anything about >32bit addressing when mmaping a file, since this impacts both the VMM but filesystems as well?
I gotta give Apple a little bit of credit, at least they have the balls to say that their transition OS isn’t fully 64-bit. This is a far cry from what Microsoft did with Windows 95, insisting that it was 32 bit. (Windows NT was the only “pure 32-bit” Microsoft OS)
Well, friends,
I will say something to who didn’t understand a thing about prices yet:
With 800,00 USD you buy EXPENSIVE machine from where I live.
Because few people can spend 800USD! This is so most places in the world.
Anything over 800 USDollars is PRO machine in most parts of THIS planet.
I wonder if it would be possible to see a release of a cheap Mobo with this G5 with integrated functions and Linux 64bits to compete with Xeon/Opteron…
You may say I am a dreamer…