First Intel Macs on track for January - Page 3

It makes no sense to me.

The one windows application I have to use is internet explorer. I'm not going to design a website in MacOSX, then boot into Windows to test, then back to OSX to develop and read my mail, then back to Windows to test my change, then back again.

That's why VPC is good for me.

The alternative for IE testing/development is to boot into Windows and stay there for the duration of my work. I'd need all my email and tools there too. I hope you can see why that's bad for MacOSX expansion.

Dual booting only makes sense for gamers where you shut off for a few hours. It makes no sense for application usage. If you rely on one particular application you must use, you'll just stay in that environment.

VPC also lets me have 3 installs of Windows in virtual partitions (and Linux partitions too for that matter). The three separate installs have different versions of IE configured. You can also wipe them out and start again in no time when they become full of guff or infested.

Also, unless Apple improves it's OpenGL implementation then I can see gamers just booting into Windows for games anyway.

I'm not impressed with that test. Very few people are.

Is it so surprising that a Quad Mac should beat a dual chip Opteron system?

Does that tell us anything interesting? Or useful?

How about comparison with quad Opteron systems as they begin to appear shortly?

Unless the Quad beat the dual opteron by 100% in every area (go back and check those tests again), it will certainly be slower.

Maybe this is Steve's five year strategy....

This has been a huge concern. While there seems to be a few on these threads who insist that it isn't the case, there is a VERY great possibility that this could happen.

The analysts that track Apple, and there are a lot of them these days, have all brought up this concern.

The problem is that if one, just one, prominent software house succeeds in dropping its Mac version, and has success selling its PC version to those Mac customers, the whole thing might snowball.

Apple has the problem of realizing that the best Windows compatibility is to have Windows running on their machine . This might make the Mac more palatable to business.

They have the nightmare of that compatibility diminishing the importance of OS X on those same machines.

Most PC people I know tell me that they get Office from work. For free.

If they buy a Mac, and can get their PC programs from work, or other places, they might feel little need do go out and buy those programs for X.

They might feel that the programs that Apple includes are enough.

Software piracy is no stranger to the Mac community, unfortunately, but it's much worse in the PC community.

Those who say that dual, or triple booting is simply too much trouble are wrong. For many people, saving hundreds of dollars trumps the extra 20 seconds or so that a multiple boot system will require.

Look at how much trouble it is to use VPC. Look at how badly it works. But a fair number of us use it when we haft to. Double booting will be much easier ,and the payback will be much greater.

Joe Shmo might not have to do much. MS makes VPC, if they want to, they can write a smaller program that will do the work for you.

Remember that most partitioning software allows you to do it without removing everything from you HD first. Apple's own software is about the only one left that doesn't.

If MS won't do this, there are plenty of shareware and freeware authors who will.

It won't be a problem.

It's highly unlikely that many apps will be ready for January. Many of those apps won't be ready in March, or April, or May, or even June.

Adobe won't make CS2 Mactel ready, for example. They've stated that. When CS3 comes out, it will be Mactel ready. That won't be until the later part of 2006. I haven't gotten any betas from Adobe yet. If anything, this is going to slow development, as both binaries have to be equivalent in features, function, and performance. By performance I mean the best they can do on their respective technologies.

I'm sure that other companies such as MS will be doing the same.

Can anyone here say that they KNOW that Apple's own pro apps will all be ready for a January launch of a PB?

I doubt it.

One thing to keep in mind in this discussion is that Apple in recent years has dramatically increased the number of apps they make in-house, from all the content apps like Final Cut and DVD Studio and Logic, to iWork and the iLife apps to Safari and iChat and cetera. Apple no longer relies on Mac developers very much, and may be perfectly happy to sell their own software to Mac users.

Yes that's the point. Any advantage in term of performance over the old ppc line, will be counterbalanced by the poverty of intelmac software.

The fastest computer on earth will be useless, if you have no software to run on it.

PIE was quoting 19 seconds for something that dual opterons were doing in 45 seconds but yes, that's just one filter and more tests are needed. Quad opterons/xeons are REALLY expensive compared to a PowerMac though and I can't see Apple using the current Xeons and unless Intel change policy and allow dual desktop chips, we're waiting for Woodcrest to get quads.

I'm really tempted with a Quad PowerMac myself now (well January at least) and even the new iMacs are looking quite tasty now they've sorted out the memory bandwidth and PCIe graphics. Core graphics run about twice as fast in some apps over the previous iMac. Dual Yonahs would be nice though if the software is available and Intel allow laptop chips in a desktop.

I agree that the quad opteron will be on par or slighty above the quad powermac G5.

However I disagree with your last point. There is very few, software or part of software, who fully take advantage of MP.
A quad G5 isn't two time faster than a dual 2,5, even if you consider the Apple benchmark (who pick up very carefully their tests ...) .
Same will apply for the quad (dual dual)core opteron vs dual core opteron.
I will add, that in general, mac os X is more MP aware, than window XP.

I commented about those tests on that thread.

I can tell you that in more controlled tests, of both PS and video editing, that Opteron systems have smashed PM's. Of course we're talking about dual machines vs other dual machines. But isn't that more meaningful?

Opteron systems have beaten an equivelent PM by over 50% in certain rendering tasks.

The timing of the first release is related to the timing of Intel's release. Apple cannot release a Mactel before Intel releases the chip that will go in it.

Because Intel is going with the dual core Yonah first the first Mactels will be related to the dual core line and the PB and iMac are perfect for these chips. When Intel releases a single core chip then the Mac mini and iBook will follow.

As for software houses dropping OS X apps, I don't see that happening. Their Mac customer base is currently on PPC and the software houses will want to continue to sell to them - that requires they keep current with both PPC and Mactel lines. Think Adobe will want to tell their PPC customer base to stuff it? No way. Same with MS - they want to keep selling Office updates to PPC users, plus VPC.

In terms of dual boot, someone has to buy a Mac before they can dual boot. That is one good way to keep Apple a computer company. All dual boot does is make the Mac computers more attractive to business. I have to use VPC to demo one small PC app - it was the only way I could make the switch. What I learned from using VPC is that you install it and the apps and then turn off internet access. Sort of like putting a condom on Windows - it can't get infected.

Oh of course. But if you can buy a quad PowerMac for about the same or just a little more than a Dual Opteron, and more importantly - it runs the right software - then why slum it with the duals. ;-)

I've 4 dual opteron PCs now running as web servers. Best bits of kit I've bought in years. Completely toast my Xeons as servers. Looking at Intel's problems on the higher end roadmap it's not encouraging for new PowerMacs so I've no problem really with a PPC PowerMac purchase. Particularly as I've an investment in PPC software already.

I don't understand why anyone is concerned about being able to install Windows on their Mac hardware -- who cares? Nobody is going to use Apple hardware as a Windows box. The real worry is that a cheap or free virtual machine is provided which can run Windows at full speed inside of MacOS X... and that developers see this as a reason not to write Mac native versions of their software.

This may not actually be a serious worry. First of all, existing Mac developers will still want to sell to the installed base for quite some time (several years at least). Second, Windows software is Windows software and Mac users have always demanded better. Third, Mac software will always be better integrated on the Mac platform (less overhead, smoother interaction with file system & devices, other apps, etc). And finally fourth, developing on the Mac has been transformed into a fairly nice (and cheap) experience with Cocoa and Xcode and for smaller developers this an important plus. The Mac software community has been growing quickly since OSX arrived.

Its not like Apple could do anything to stop virtual machines from being created for OSX anyhow. With an x86 processor and Darwin being open source, a Windows VM is inevitable. The real question is: how cheap will it be? My guess is that it'll be free, you just need to buy a copy of Windows.

We are, of course, ONLY talking about MP aware, ot mutliply threaded apps.

I wouldn't agree that X is better inthis area that Windows. I think it's a toss-up.

When Vista and Leopard arrive it will be more interesting. XP, at this point in time, is a lame duck. It should have gone two years ago.

Of course, the point I was making is still valid because the Opterons have a much better memory system than the G5's. The onboard controller gives them a big boost up over the Mac's well known memory latency problems. The fact that Apple is so conservative with its memory specs doesn't help.

If Apple had gone with DDR2 667 RAM instead of the 533, it would have made a big difference. The lack of advantage of DDR@ 533 over DDR 400 is well known. The best that can be said for it here is that in Apple's design, it would help "somewhat". IBM finally added to the caches. But it's been said that the 970 really needs 2MB per core, especcially in a system where two cores now share one memory bus. Cache adds little to power consumption, but helps reduce the effects of the poor latency problems we see.

So onboard controllers plus 4 channels to external RAM and bigger caches, will help a quad Opteron system to outperform the Mac Quad.

in terms of software apple should make the transition as long as possible...
if half of the macs are PPC and half intel, the developpers will have to go to universal binaries.
Once they have launched their software in universal binaries, most of the work is done and they will continue to develop for OSX.

And the best think would be PM/iMac stay in PPC G5 for at least 2 years from now, and the other go to intel. so even pro apps (PS, Quark Xpress, etc..) will have to switch to UB..

Not that great for 3rd parties or future software sales though.


Fine if the app you need isn't an internet based app. Unfortunately for me it's Internet Explorer I need VPC for - perhaps the worst internet app I could possibly need to test. However, it's pretty well contained though. If you do get infected you just delete the VPC session and create a new one. I've a Windows2000 disk image I mount and install from so it takes no time at all.

are games really THAT important? maybe i'm missing something here

That's quite true, but Wine and Crossfire (I forget the exact name) can't run the full range of Windows programs. If a program doesn't fit within the API's they have already brought over, more work has to be done.

While this is fine, as many programs will already work, it's not seamless. You have to check first.

And as you say; "this MAY to be a worry".

I'm hoping that it won't be. But it might.

me to. haha i didn't even know there was a "mac" gaming community.

how many people does that refer to? a few hundred thousand? if even.

That's what I meant. For people like you, it doesn't make sense, but the average user is not a web developer, or application tester or what have you. There are many, many more gamers out there, and I see dual booting as a solution for them. Then you get the best of both worlds. I had a gaming PC until recently that I'd switch back and forth from. I got rid of it because it was too loud and I'm not much of a gamer anymore. But if I could game on the same machine that runs Mac OS X, I certainly would, as long as the performance of said machine is somewhat comparable to a standalone PC. That may not be a possibility.

Some people are saying that they are going to wait for the second wave of intel macs to work out all the kinks.

What kinks will there be? Theres nothing unusual about the intel processors, theres nothing really different.

and apple have been monitoring its os on intel for 5 years.

Well, you just answered the question.


This may be true for developers that write software exclusively or mainly for the Macintosh. But the bigger ones who develop for MacOS and Windows, I am afraid that they will happily drop the Mac versions once they realize that their customers can run the Windows counterpart of their software in native (or near to that) speeds on an Intel Mac. And this is going to happen since many people have or can acquire Windows software at lower or, in some cases, not at all cost. You just cannot underestimate the corruption power of this fact for the Mac software.

Saying that this may not be a serious worry is, for the time being, just a wish and hope.

It is not just the processor. It is going to be a completely new architecture. Apple needs to make sure that the OS plays well with all the system components. From the lowest to the highest level.

I don't think Apple has any idea how long third parties will take to transition to Intel. They're in a chicken-and-egg scenario where developers are waiting for a market for their apps and Apple can't release hardware without any software to run on them.

We won't know how this will play out until it does. All we know is the sooner its done and over with the better (especially to replace G4-based Mac models like the PowerBook and Mini).

If Apple couldn't release Intel Powerbooks until January and could release new PowerPC models before Christmas, then their course of action is easily understood. I fail to see any irony.

Two general comments,

1) It seems smart (to me) for Apple to release MacTel's ASAP to spur developers to port their code. Why? Because now there will be tangable HW, no excuses left for foot dragging by developers who won't start UNTIL there is existing HW. It would at least seem to be prudent of Apple to invest heavily by deploying their SW engineers to various developers to insure a rapid transition.

2) In the past Apple has been very secretive about product releases (look at the quad G5 specs. (the actual model specs. (such as CPU/bus/GPU speeds, etcetera), which NO rumors site obtained (not even TS)). I've been thinking for some time now that with all the existing SW that must be recompiled for MacTel's, many, many 3rd party developers will be in the loop as HW products are developed. It's all about getting SW developers to port their codes ASAP. I don't think Apple can keep a lid on this one, and I don't think that Apple even wants to. I even think that this may be the one time when Apple wants people to know what their up to. Look, see, were making progress. Look, see, were 6 months ahead of schedule. No, I think there will be a steady release of "rumors" and maybe even that there are people within Apple specifically assigned to "leak" said "rumors."

Do either of these "brain frats" make any sense?

As a related aside, Apple has posted a new edition of their Universal Binary Programming Guidelines, dated this past Wednesday:

http://developer.apple.com/documenta...ary/index.html

They seem to be revising it based on actual scenarios people may be having issues with, and adding tips for helping the port go faster. Several of the new "tips" related to CodeWarrior developers, as well as revisions for XCode 2.2.

Judging by some of the additions to this document, a lot of developers have been giving them a lot of feedback while porting, which I take as a good sign.

"It seems smart (to me) for Apple to release MacTel's ASAP to spur developers to port their code"

With what a utter joke Intel's chip plans look like right now for the next five year and how fantastic the PPC systems are that Apple is shipping, Apple better be thinking of how do they get IBM to change their mind about not wanting Apple as a customer anymore.

And there is still plenty of time between now and next year for even more Intel cancelation and slippage. It's too depressing to even read x86 sites anymore about Intel and think that is what Apple is going to be stuck with for the next five years.

I think it's a waste of time to wait for the "second wave".

Apple just went over to the Express bus. This is quite a major change. I'm waiting until January to buy a Quad, because I want to see what small problems will arise in the first 25,000 or so. But I'm not thinking of waiting for the second wave, even, at this point, if there would be one.

When the Mactels come out, there will be some small niggling problems as well. Nothing major. Apple will have had plenty of time with the hardware and software. But any new technology a company is dealing with will have teething problems.

But those problems should be minor, and cleared up after a month or so, at the latest.

It isn't something I would worry about.

I think they have more than a fair idea about most. Avid just said that they will be moving their new software to Universal binaries. The new PPC software will be available mid 2006, and the Intel versions (MAC) will be ready when the machines are out.

I don't think Apple just guesses. They talk to their developers as well.

It depends on how far from the Intel platformisation product Apple have drifted. I can't see them being too different.

One sure way to do that is to drop some heavy cash on IBM's CPU development table. Apple does not seem very inclined to do so.


Never heard of the Apple Curse(tm) before ?

I can't agree with that.

First of all, no chip maker has a roadmap for the next five years. What they have is a tentitive plan for the general direction as well as for the technology they would like to use.

The roadmap extends across 18 months to two years. Past that it becomes hazy. That's why the 90nm roadmap for all of the makers didn't work out. Unexpected problems hit all of them at once.

Intel's roadmap was completely washed out, as they depended most upon the technology that was most affected by 90 nm's (and beyond) problems.

But they made an amazing recovery. Their roadmap looks pretty good. They are even moving production of chips not expected until late next year into the March timeframe. No one else has the ability to do that.

Right now, we have no idea as to what IBM's roadmap is. Is that better for you? Can you tell us what IBM has planned for the G5, or for whatever comes after? What speed bumps is IBM planning, and when will we have them? How much more cache are they planning to use? When will they go to a consolidated cache for the dual chips? Will we see an L3? Are they planning an integrated memory controller? If so, when should we expect it? What are they planning about low power chips other than the one they announced? Both AMD and Intel have discussed plans for quad cores in 2007, will IBM be releasing quad cores then as well? What will they be like? How will their compilers take advantage of them? What about memory bus issues with quad cores?

Etc., etc,.

They'll be replacements. This is a transition, not an addition. By replacing them, it drive developers to move to Xcode and port their apps to Intel. As long as developers stick with Xcode, their code portability will be comparatively very good.

In terms of performance per clock rate, in depends on the application. In integer apps, a Yonah core will be faster than a 970fx. In FPU apps, a 970fx will be faster than a Yonah core. For SIMD, any thing permute related, the 970fx will be better, otherwise about the same.

In general a Yonah iMac will feel faster than an iMac G5. These are things involving the GUI and filesystem. For compute bound apps, it'll breakdown like in the above paragraph: Yonah wins in integer, 970fx wins FPU, 970fx marginally wins SIMD. What trumps the 970fx is of course Yonah is a dual-core chip with less power consumption than a 970fx at the same clock rate. That's a very big win.

Intel will also make a few Yonah "derivatives":

1. "normal" for regular laptops (up to 2.2 GHz)
2. "low power" for more mobile laptops (likely <1.8 GHz)
3. "ultra low power" for small laptops and tablets (likely <1.6 GHz)
4. "extreme" or something like it for gaming (likely 2.4 GHz at 50W)
5. "Sossaman" for blade and 1U servers with multiprocessor (4-core) chipset support (probably identical to number 4)
6. "single core" versions for cheap <$900 laptops
7. "Celeron M" versions which may or may not be single core and may or may not have 2 MB cache in 2H 06 or later for <$900 laptops

My guess is that Apple will use the 1.8 GHz low power dual-core variant for Powerbook, if in fact they are ~20% thinner - that's a 0.8 inch or 20 mm thick laptop with all of the goodies, wowsers! - and the normal variant for a 2.2 GHz dual-core iMac. Both of these would be considerable upgrades from the previous PPC machines. Both of these would be able to play 1080p H.264 movies at full framerate. Well assuming the Powerbook has 2nd HD capable display and the iMac has an HD capable display.

That's pretty close.

The only things to add to that is that the Yonahs are planned for a max 2.33GHZ speed, and they have 2MB cache per core, as opposed to 512Kb for the 7447a, and 1MB for the so far imaginary 7448.

So you figure that software for the Mac will meet its demise? Or that it is a likely possibility? What will be the percentage in Apple doing this switch, to be able to sell more computers?

You are referring chiefly to Adobe, Macromedia(oh oh, no more) and Micosoft then, or to Avid too?

That would leave smaller companies who develop for the Apple platform, and Apple left to develop software for the Mac.

That I disagree with -- Intel and AMD haven't shipped anything yet that has comparable throughput to the 970's VMX unit, even when at a notably higher clockrate (in the case of the P4). Between the small register set, lack of fused ops, and halved instruction through put, the SSE units just don't stand up to the VMX units unless the hardware you are looking at has significantly better bandwidth (and the G5s do pretty well in terms of achieveable bandwidth).

Actually I was just refering to installing Windows on an Intel Mac as if it were a PC in a dual boot configuration. There is no reason to worry about people doing that at all.

The Windows-in-a-VM thing might be... but I don't think so. The developers who are likely to drop Mac development because of it are going to be the ones doing a louzy job of Mac development in the first place, so it won't be too much of a loss. Hell, their software might even work better under a VM. There will still be a market for quality native Mac software, and this market is likely to grow over the next couple of years while the importance of supporting the PPC hardware is still foremost in developer's minds. Remember, it is likely to take 5 or more years before the x86 installed base outweighs the PPC installed base.

Would it really be a waste of time? Seems to me there's quite a bit of interest in this thread for running Windows on a Mactel. But the discussion seems to focus on current Windows, i.e. XP. It's a fair bet that the 2006 model Mactels won't be able to run Vista optimally. The massive 2GB RAM requirement for 64-bit Vista doesn't help there. So early adopters will likely be limited to XP, which I guess isn't bad, but I would expect the bleeding edge wouldn't like to be trapped with an EOL OS.

While that may be true, I wouldn't count XP out just yet. I know quite a few people (myself included) that still run Windows 2000 reliably and will for sometime to come. I'm sure if Vista doesn't work optimally, as you say, XP will more than suffice for a number of years. If you want to be bleeding edge as far as PC software/operating systems go, you can always build yourself a bleeding edge PC.