Image title: Apple and Intel
From: jeffandcaryn’s Flickr album.
I know many will take offense to being called a fanboy. Well, too bad. If you are spouting forth woe and misery based upon hearing Apple will be using Intel processors in the future, you are a fanboy… deal with it.
Aside from the fanboy bashing, I wanted to share my thoughts on yesterday’s keynote address. Well, I will ignore all the marketing jargon in the beginning which spoke of how well the Apple retail stores are doing, how well the iPod is doing, etc.
Let’s fast forward to when Steve uttered the word: Transition.
As Steve explained, Apple has undergone a couple major transitions in its lifetime. The first transition involved them first moving to the PowerPC architecture. Steve seemed to think this was a good move, although he did feel the need to mention that he wasn’t around then. Sounded a little like, “Hey guys, it’s not my fault!”
The second transition alluded to was the one that occured about 5 years ago, when Apple left behind OS9 and moved ahead to the radically different OSX. Apparently, the process of porting applications was exceedingly difficult.
The next transition, as we all know now, is for Apple to say goodbye to the PowerPC architecture and start adopting Intel x86 processors in its computers. Only Apple can Think Different by moving to the architecture that the majority of computer manufacturers already use. But, all jokes aside, this is a radical move. This is the sort of thing that not many companies would have the stomach for or the gonads to attempt.
Steve is a showman, and his keynotes are often met with enthusiastic applause, but despite the press coverage for days speculating about this move the audience greeted the “It’s true” presentation slide with stunned silence.
The audience started to come around a little when Steve showed them that all the demonstrations and the presentation itself was running on an Intel Mac. Not too shabby. Oh, and when he spoke of the secret life of OSX in a manner in which only Fox Mulder could duplicate, the audience began to soften and relax.
Then came the demonstration from the folks that produce Mathematica. Apparently, this “beast” of an application package was tweaked and recompiled for the Intel-based Mac in only two-hours. The main kicker is that it only involved changing approximately 20 lines of code… out of millions.
The crowd, surprisingly, became most enthusiastic duing Paul Otellini’s (CEO, Intel) speech. His showing of the “Apple toasts the Pentium” commercial broke the ice. And, he stated that Intel didn’t hold a grudge (yea, right).
When all was said and done, the Developers seemed to be fairly excited about the switch. Early reports even suggest that many think it is the right move (despite the extra work that will be required to port existing applications).
Now, if only we could get the Apple fanboys to shut-up and stop declaring Armageddon.
What the fanboys likely don’t remember is the keynote speech Jobs gave in 1997, when he accepted a 100 million dollar buy-in from Microsoft to save a floundering Apple Computers. So it isn’t exactly an unprecedented move.
If anything, fanboys should look forward to Apple using the faster Intel chips. The only thing I’m waiting to see is how the OS will run on CISC architecture with 16 registers (for the 64-bit CPU) or 8 registers (for the 32-bit). Faster clock speeds or not, that could be a bottleneck.
So here’s the really important questions:
1) When can I get a copy to install on my existing pc?
2) Can I play windows games on it? Or will it at least make it fairly likely that OSX ports will happen some time within our lifetime for most games?
I am not sure how the MS deal has anything to do with this move to Intel… but, whatever.
As far as how it runs, they have stated that they explicitly designed OSX to be architecture independent, which should make the transition for the OS fairly easy. Did you even watch the keynote? OSX looked pretty snappy on the P4 3.6GHz machine. Since they have been maintaining this “Just in case” build of OSX for the past 5 years (not to mention every piece of sofwtare they have written), most of their work is already done. Most likely the remaining work will involve optimizing the OS.
With regards to OSX on an existing PC, Apple has stated they will not allow it to be installed on anything other than Apple hardware. No word on how they will implement this. If you read the EULA for OSX, it alredy states that it can only be run on Apple labelled computers.
Ahh… Gaming. That will be interesting. Virtual PC will become unnecessary, and will be replaced with something closely resembling WINE (I suspect). This should mean much better performance for Windows software in OSX. However, in the same statement regarding OSX on non-Apple computers, Apple stated that they would do nothing to prevent someone from running Windows on an Apple computer. Dual-booting for games would certainly be an option for many (future) Mac owners.
IANAD, but I think one of the larger obstacles to porting games from Windows to Apple has been DirectX over OpenGL. So, I am not sure if porting becomes any easier.
As for OSX on non-Apple hardware, they’ll probably put a simple “i’m an apple” chip on the motherboard (moto does something similar on their blades), and check for it on boot. I remain fully confident that haxx0rs will release apple-check-free versions quickly after every official release. Damn lazy Apple, not wanting to test on the mad variety of PC hardware…
Comments on the announcement from people actually involved in mac games development.
Mathematica has always been cross-platform, so no kidding it was easy to port, duh.
But there’s no such thing as an architecture independent operating system, for obvious reasons! Although the Mach microkernel was probably a heck of a lot easier to port than, say, the linux kernel even. So maintaining an Intel build of OS X was probably not something they had to spend a lot of development resources on (remember that NeXTSTEP was originally written to run on x86 in the first place; I’m sure Jobs always understood what a handy thing that was to have).
As for games, no, you can’t play Windows games on it. C’mon Dave, you know how these things work! At least, you know as well as I do. It’s not just going to be matter of recompiling the game’s object code to port to OS X, and it won’t even make it significantly easier to port games; as Colin rightly speculates, the major hurdle is usually DirectX, but games have all the same problems that other cross-platform programs do.
I doubt the switch will get more games on the Mac platform, unless it results in a large influx of new users that are gamers. With game bugdets spiraling out of control, publishers are increasingly chasing the dollar signs, and spending money on an OS X port probably isn’t a break-even proposition for most games. Of course, I’m not a game developer, so I can’t say for sure that this is the case.
Virtual PC will remain necessary — WINE has never really worked as well as VMWare, for example — but not needing to emulate the x86 instruction set on a PPC will make Virtual PC a heck of a lot faster.
Lastly, I think Rosetta is a losing proposition. They’ll get 60% of the performance that a native PPC app would get on actual the PPC architecture, if they’re talented and lucky as all hell. Emulation of PPC on an x86 processor is, as I’ve already said in another comment, non-trivial. It might be good enough to run a few clunky old applications that you don’t use much, but you’ll want a native version of any applicaiton that you expect to run any faster at all on your brand new Intel Mac. How many people thing that Adobe, Wolfram Research, Quartz, etc., are going to be handing out free upgrades of their $1000+ per seat applications for free? I fully expect the development companies to be very happy with this announcement, since it’s the biggest cash cow to hit the mac platform since the original switch to PPC from the old Motorola 68000s.
Carl, I’m not really sure what you’re trying to say here, but it doesn’t seem to make much sense — Apple has maintained a version of OS X that runs natively on x86 hardware all along. None of the operating system’s object code or the bundled applications will be emulated, so the number of registers on the Pentium 4 won’t have any impact on the performance of the applications at all*. This is what compilers are for, after all: optimising object code for specific architectures.
Also, it’s not clear to me that the Intel chips are all that much faster. The G5 compares very favorably to the Pentium 4. You can find the benchmarks online, if you care to look. I think the real reason for the switch is that Apple is getting a better deal from Intel then they ever got from IBM, so it will improve the margin on every mac that they sell. Jobs did a great job selling it to the crowds at WWDC, but Apple’s real motive isn’t making the developers happy, it’s improving their bottom line. Something that I completely understand — as Carl points out, it was only a few years ago that people were predicting the “end of Apple”, their finances were in such dire straits.
* At least, it will impact them to the same degree that it impacts any program that runs on x86 hardware, from Microsoft Word to Doom 3.
True. But the key isn’t currnet PowerPC products v. current Intel products. This move is all about what is happening 1-2yrs from now and beyond. IBM has little motivation to spend resources improving their desktop offerings from Apple, whereas Intel apparently has a clear roadmap for the future. I am pretty sure the dual-core Pentium M processors that people predict will be in the PowerBooks 1 year from now will outperform the current top-of-the-line PowerBook.
Right. But, I think it is more a reference to a development mindset. It seems that Apple intended to keep OSX as easy to port as possible, and the fact that they were maintaining builds on two separate architectures for the entire life of the OS tells me that this isn’t a by-product of anything but rather very intentional. Maybe they shouldn’t (and I shouldn’t) say architecture independent, but maybe architecture flexible.