Subscribe to
Posts
Comments
NSLog(); Header Image

QotD: Intel

Question: Do you agree or disagree with Apple's decision to move to Intel?

My Answer: It's too early for me to tell, as I've not read much about it, but I'm curious what y'all might have to say.

You are encouraged to answer the Question of the Day for yourself in the comments or on your blog.

18 Responses to "QotD: Intel"

  1. No, not really. You can check out my opinion/write-up at http://vinay.afternight.org/index.php?id=466 (temporary housing for my blog while I move vinayvenkatesh.com)

  2. The whole move is about Notebook chips. Not the 3GHz bit, IMHO (Not to say it doesn't come into play, it's just not a big factor). The majority of non-iPod Apple hardware sales are notebooks. The goal is to move to a company with a dedicated research group on notebook chips. Neither Freescale or IBM has this, AFAIK.

    I think the move in the long run will be for the better. However, I think the ride will be a little bumpy.

  3. I'm hopeful that it goes real smooth, I think it will be just about as smooth as a transition like this can be. Thankfully the transition fits right into my regular upgrade cycle of new hardware.

    It's definitely all about the portable more so than the desktop, but it's not like the desktop will be suffering for it. I think a lot of people mainstream people miss that Apple is more of a hardware company than a software company. The Mac will still be a better computing experience than Wintel, Apple can still control the experience with a x86 processor as a PowerPC. I don't drink the PowerPC kool-aid I drink the whole Apple Mac OS X kool-aid, what drives that experience as far as the CPU matters little other than that it works, it's fast and isn't power hungry. Being a mostly laptop person I'm happy to see Apple make a strong move now to do their best for their hardware in the long term. I also like how someone put that its best Apple does it now in a position of strength instead of weakness.

  4. You've seen some big developers be relatively quiet the past few months, and I have already abandoned even wanting to deal with software that is not completely Cocoa at least. Lets just hope we will not have to use Windows to use the Adobe/Macromedia suite.

    I'm taking things optimistically- I wasn't expecting the multi-core Dual processor fast G5 I WANT to have anyway, to be available this year. The Intel that will end up in Macs will be the better Intel, not the failed models talked of in the pre-longhorn moratorium. Good Laptop Chips, Pentium Ms, Good multicore 64 bit designs from the DEC Alpha tean, like the Tanglewood.

    And probably, increased game development and Media availability. Better Console and SetTop box and Video Card possibilities, and run Windows niche software, if you have to, at native speeds. ITunes Movie stores. Universal Game development platforms.

    And at the very least, a chip manufacturer that can produce yields that a customer needs, without having their loyalties torn by other customers.

  5. Been there, done that - if Rosetta works well enough, this is about as painless to the user as the 68k transition. Once people have made their code cross-platform with universal binaries (you can do that with your software today by downloading XCode 2.1), Apple can choose processors to its heart's content: they can move from Intel to AMD if they make no progress, they can also adopt new PPC chips if they are attractive again. If the transition is really what Steve Jobs said it is, it will be easy enough to develop apps with both CPUs in mind.

    With that advantage, I'll gladly sacrifice some performance points for the portability of a Mach-based OS.

    A quick summary for you, Erik (at least those points were important to me, you will read it up anyway, so this might well be redundant, sorry if it is):

    the transition period will last two years - through 2006/07, we will see PPC and Intel Macs

    Apple will offer a transition SDK - i.e. an Intel-based Mac - to ADC Premier and Select members for USD 999 within two weeks from now

    the Intel Macs will have a seamless translation software called Rosetta for legacy PPC apps

    Phil Schiller said Windows might run - Apple will neither stop nor support it

    Apple will not allow OS X to run on stock peecees

    XCode 2.1 is available as GM via ADC, install the 10.4u SDK and you can start building universal binaries

  6. I haven't read a whole lot on the transition yet, so I might be missing some details. Having said that, here you go:

    If they use this as an excuse to quit making hardware (think "NeXT"), I think this is a terrible move. Conversely, as long as they keep making hardware, I don't really care what the CPU is -- as long as Apple doesn't start putting "Intel Inside" stickers on our machines!

    It's pretty clear now why there was such a big push to get to OS X and Xcode. I bet this was Steve's plan all along.

    Oh, there is one other thing that concerns me: OPENSTEP was ported to m68k and x86. Rhapsody was ported to both PPC and x86. x86 is 32-bit, and unless Jobs and Co. can get that 32-bit x86-based OS X version ported to Itanium within a year (I doubt it) then we'll be stuck with 32-bit x86 hardware next year. And unless they are already planning on ANOTHER transition (to Itanium) after that, this move seems like something of a step backward from 64-bit G5s.

  7. I pretty much agree with John Siracusa's take on the whole situation: http://arstechnica.com/columns/mac/mac-20050607.ars

    I don't like the x86 ISA, in fact I hate it. It's a bloated mess. But if this means Apple will be able to produce faster mobiles and great future machines, then so be it. (Although I do not look forward to the day I'll be staring at x86 instructions in Xcode's debugger instead of PowerPC.)

  8. IBM said that at no point was anything more than 2% of their facility dedicated to producing chips for Apple. I think i can understand why apple is on the move. I'm sad, scared, etc, but i think it'll end up being a pretty powerful move.

    At least apple's doing it now, from a powerful standpoint, instead of in 10 years, from a desperate standpoint. These are strange times, though.

    I'm really just surprised that it's x86, and not some new technology. I would think Apple would make the switch when Intel starts making something better than x86, but i would be wrong. 🙁

  9. The x86 instruction set truly is horrible, but almost nobody works at the machine-code/assembly level. Apple just can't compete with Intel's huge engineering resources devoted to cranking out faster chips.

    Of all the things I've heard, these seem the most important:

    - To do a movie version of itunes, and to support H.264 HDTV format, Apple really needs the CPU power.

    - People love Apple for reasons other than the PowerPC instruction set.

    - Given the same Ghz, PowerPC is slower than Intel.

    - Given the same Ghz, PowerPC is hotter than Intel.

    - Given the same Ghz, PowerPC is bigger than Intel.

    - Each of these can be addressed, but only by making the product more expensive.

    Insofar as the only important thing is what the enduser experiences on his monitor, Steve Jobs made a good decision. Without switching to Intel, Apple was set to lose the $1000 Apple Laptop vs. $1000 WinTel Laptop competition bigtime.

  10. @shredstar: have you ever wondered that the exact opposite about PPC vs. Intel has been stated by many MacHeads a gazillion times before? BTW, on a "given the same GHz" basis, I still believe the current G5 lineup leaves the P4 in the dust, bigtime. OTOH, if you are referring to the Centrino vs. the G4, you are right.

  11. I personally don't quite understand the move...

    So when Intel reveals their next Pentium next year and Apple shoehorns it in a Mac, we will have a Mac that is 'hopefully' as fast as a Dell running Windows using the same chip. But the Dell will, probably, still be cheaper (industrial design, rest of components, mass production, etc).

    So we will end up with a machine, that is, at best, as fast as the fastest Windows PC but more expensive. Better looking of course.

    Or, Apple could use the motherboarddesign of some PC manufacturer and try to build Macs really cheap. But we will end up with BIOS (According to Apple's documentation the IntelMac does not use OpenFirmware...).

    Hopefully they will keep the option of running some version of the MacOS on legacy Intel/AMD hardware, and use it in the future.

    Or maybe release a yellowbox environment for win32, so developers can deploy their x86 code on another userbase as well.

  12. Daniel: just a minor correction - NeXT OS was available for Moto 68k, Intel X86, Sparc, and HP-PA RISC systems. 4 way compilation of our apps (both custom and commercial) was via checkbox just as with the upcoming Mac OS, and it worked absolutely seamlessly. Being cheap, I never equipped the office with PA-RISC systems, but we released across all 4 and never ran into an issue with any of them - it all compiled the first time, and went through beta and into release without any architecture issues at all.

    -M

  13. Oops - bad karma for a double post here, but to address Zoefsnoep's question:

    So when Intel reveals their next Pentium next year and Apple shoehorns it in a Mac, we will have a Mac that is 'hopefully' as fast as a Dell running Windows using the same chip.

    I'm quite confident that you'll end up with a Mac that's significantly faster than a Dell running on the same chip. Windows has a lot of cruft, their architecture and API's all contribute to a much slower software experience than could be had on the hardware it's all running on.

    Back in ancient history during the Yellow Box days, when we ported our Cocoa (OPENSTEP) app over to Windows, the whole office was just floored when we found that our app, using a lot of the architectural goodness you'll see in OS X for Intel including double-buffered graphics and a "window server" absolutely demolished the native Windows equivalent in speed.

    It wasn't just 50% faster. It was more like 5-6 times faster. Now, that sort of architectural difference might start to be minimized with Longhorn and it's Avalon windowing environment, but Windows developers are going to be spending years moving over to the new Windows stuff and the old cruft will persist.

    I'm going to go out on a limb here and state that a lot of the speed you see in Mac OS today has been bought through the toil of Apple first in optimizing its own code (think of the general speed ups over OSX releases) for what has always been a more performance-constrained PPC environment, and by developers leveraging that toil and adding their own 🙂

    One thing that we noticed back in the day when we moved from 68k to Intel, was that Intel's integer performance (versus 68k's and PPC's great floating point performance) is used to great effect in making the OS run fast. That work that Apple has done over the years to make each iteration of OS X faster than the previous one will very likely show up on the Intel side and act very favorably for the performance of even the first shipping Mac OS X86 PC. The fact that Apple has been quietly working on this all in parallel to OS X development is good reason to expect big things when the switch happens.

  14. Initially I was a bit bummed about it. PPC instructions are more elegant than x86 (I don't think anyone will disagree), but x86 chips have a roadmap that has faster machines. After my initial shock, I have to say I am fine with the switch (though, I still feel a bit dirty for saying that, hopefully that'll pass). I went through the 68k to PPC migration and it was utterly painless. I'm confident that the PPC to x86 migration will go just as well, if not better. Hopefully I can pick up some Dual G5s in the bargain bin. 🙂

    Oh, one thing I won't like is little endian. I hate debugging little endian memory.

  15. I am hurt to the bones over the emotional facts et al ..

    But I have spent considerable time studying the upcoming Yonah and Conroe Chips, and what the Intel Roadmap has even further ahead.

    Steve said, its was something he saw in intels future (looking at roadmaps together with pat?), so, if this and the lack of a suitable g5 powerbook has gotten him into this decision, i can understand.

    Apple has been struggeling for CPUs for so long I almost forgot when they actually had NOT to ...

    I like Yonah I have to admit. Its not a really exceptionally geeky CPU (all really cool CPUs are dead), but its a good one.

    We are NOT talking NetBurst architecture here in terms of Apple's move!

  16. I'm quite excited about getting my hands on a wicked-fast Powerbook, I'll have to admit. But I am quite concerned about the PowerMac which, in truth, is more important to me and how productive I am.

    I'm curious why there doesn't seem to be much discussion regarding the 64-bit XEON MP, which seems to be much more comparable to the G5 than the P4. Apple has really marketing the PowerMac as a workstation since the G5 and OS X and I think the P4 is anything but a workstation-class processor, regardless of what HP and Dell think.

  17. The Xeon is probably on a dead branch of the Intel growth tree. It is not Intel's future for 64 bit or multiple core, and certainly has its own history of struggles.

    The P4 of the development kit is not the intel mac of the future, it is merely to assure clean stable code. You might find its relatives more directly in a Mac Mini.

  18. Mac on Intel: Roundup and Reactions

    In case you've been living under a rock for the past 24 hours, Apple has announced that they will be...