Behind the Apple-Intel Deal, IBM Surprised

[VanNess]

I told them the change won't really affect them, and that they should just go out and buy the computer they want now and not worry about the Intel change.

But, the fact that this was a concern by people who are not very computer savvy and who don't read sites like this or have a subscription to Macworld leads me to believe other "normal" computer buyers may have the same reservations, which could hurt Mac sales this year.

I think you made the right recommendation to your friends.

If there are any potential upgraders/switchers who have concerns (or more likely, questions) right now, it's mostly due to the somewhat ambiguous media attention all of this has gotten. All that's really known for certain as of today is that Apple, in the future, will begin using Intel possessors, and that's it. The vacuum thats left by not having any other concrete information seems to create a psychological need on the part of the media to fill in the blanks themselves, so the good old-fashioned guessing game machinery springs into action. That results in a fair amount of conflicting conclusions and general confusion for people who aren't techno-savvy. (and based on some of the stuff I've seen, even some techno-savvy people who should know better - hello Bob Cringley 🙂 ) But it will all fade away into the background soon enough, until there is something further to report, and that will most likely be the introduction of the first Intel-based Mac. In the meantime, new iPods, new Mac software, new gossip (Apple and the potential movie download business) will relegate the attention about the Intel switch to geek message boards arguing forever in indecipherable technobabble (and more guesswork) that won't have much meaning or impact for general users or new computer buyers.

When Apple ships it's first Intel based Mac, and it becomes clear that nothing at all has changed for the end user except performance improvements that would otherwise be expected for a product refresh, then the attention will almost universally shift to Leopard and Longhorn predictions if it hasn't already by then.

 

[VanNess]

I have my doubts those chips are contracted for a further two years.

I haven't verified this, but according to a post I saw on another board, the IBM agreement is set to expire on 10/07. That dovetails with Apple's hardware transition timeline and Jobs statement that there are further PPC based products still in the pipeline.

The post cited Apple's disclosures in it's 2003 SEC filing. The contract was signed in late 2002, If I recall correctly.

 

[MontyZ]

Yes, it's the vague way the media has reported this story, as though Apple will introduce all new computers with Intel chips. It could make people think they should wait until then before buying a new Mac, and that's been my only worry with the Intel announcement. I, personally, don't care what chip Apple uses, as long as it makes the computer fast and reliable.

Let's hope the new Apple ads coming out will dispell some of the inaccurate claims and rumors.

 

[VanNess]

Yes, it's the vague way the media has reported this story, as though Apple will introduce all new computers with Intel chips. It could make people think they should wait until then before buying a new Mac, and that's been my only worry with the Intel announcement. I, personally, don't care what chip Apple uses, as long as it makes the computer fast and reliable.

Let's hope the new Apple ads coming out will dispell some of the inaccurate claims and rumors.

I don't think Apple should do anything marketing-wise about the Intel switch, at least not yet. This is going to burn out pretty soon, as it should. All in all, there simply isn't anything concrete for the media to write about right now, and that will become apparent. Maybe it already has. If Apple says anything further now, or in the near future, short of a Intel product introduction, it will just fuel the fire all over again.

 

[MontyZ]

I don't think Apple should do anything marketing-wise about the Intel switch, at least not yet. This is going to burn out pretty soon, as it should. All in all, there simply isn't anything concrete for the media to write about right now, and that will become apparent. Maybe it already has. If Apple says anything further now, or in the near future, short of a Intel product introduction, it will just fuel the fire all over again.

I'm actually more concerned about how Apple's rivals will use this to spread misinformation in an effort to disuade people from buying a new Mac. But, we'll just have to wait and see.

 

[MagnusDredd]

How is a PowerPC with 8 SPE's less effective then a PowerPC w/o 8 SPES ?

One more time.

The PPC (970) core in the Cell CPU is terribly stripped down. They have removed many units which make the CPU able to run normal programs with any speed at all. This includes most everything that currently runs on OSX. From what I read some apps would simply not run at all.

Some applications _could_ be re-written to take advantage of the Cell, Photoshop being one of those. To rework Photoshop to the degree required to take advantage of the Cell, would take more time than it's worth. This is because the Cell requires a completely different approach to programming. And even if Adobe spent the tons of money and time rewriting their apps required (which would cost more than they'd make off of them), everything else would still be very slow.

The Cell is an incredible embedded CPU. It is a horrible general purpose CPU. The only use I can see for it in a computer would be on an add-in processing board (sort of like the blue ice boards). Below is a link to page selling a blue ice board:

http://www.bhphotovideo.com/bnh/controller/home;jsessionid=CwjhHnwTjR!-132689729?O=productlist&A=details&Q=&sku=184839&is=REG

Note: I'm pasting this message into stickies since some people seem to be too slow to get it the first time.

Also: The Xenon processor in the XBox has cores that are as stripped down as the Cell. There has been some speculation that the Xbox 360 won't be any better than the current model for running AI and physics. Which means the games will be much prettier, but the AI may be as stupid as it currently is. The same applies to physics. I'm not sure how much I would expect as far as advancement of game physics for the XBox 360 or PS3.

To put this simply, the Cell and Xenon are terrible choices for computers, but pretty cool for specialty stuff.

 

[MagnusDredd]

As it stands, peripheral support from some companies is lackluster/shakey at best when it comes to the Mac. I can't help but wonder how many manufacturers will decide this move to Intel is the final straw and decide to drop Mac support for their peripherals and/or devices? Anyone have an idea how much work will be required for drivers to work with OS-X on Intel?

Probably not much.... Remember OSX's driver structure was designed for this very thing. NeXTStep ran on multiple architectures simultaneously without issues.

I actually think it'll increase driver support. For all a Mac is, it's still just another PC architecture. The PCI slots work the same. This is why I can put the same network card into my Mac that I run in my x86 boxes under Slackware and XP. USB is basically USB.

The problem with driver support has also been one of perception. People (programmers included) still have this idea that Macs are entirely different, and need special cards, etc, etc. This is really only the case for video cards and drive controllers. The move to Intel chips will push this point home for many people. The only reason most PCI cards don't work in Macs is lack of drivers/software.

 

[steeldrivingjon]

Some applications _could_ be re-written to take advantage of the Cell, Photoshop being one of those. To rework Photoshop to the degree required to take advantage of the Cell, would take more time than it's worth.

It would probably work best if a Cell processor were used as a coprocessor, alongside a regular PowerPC (or two). Even if Photoshop could be made to work on a Cell-based computer, it'd be kinda useless without everything else a user relies on.

Photoshop on a Cell would be kind of like the old Pixar Image Computer, which had its own proprietary processor tuned for the kind of graphics stuff Pixar needs to do. (They got rid of their hardware a long time ago.) It was good at graphics but it wasn't useful for much else.

But, in the end, if you're going to use a souped-up chip as a graphics coprocessor, you might as well just use the GPU.

 

[steeldrivingjon]

People (programmers included) still have this idea that Macs are entirely different, and need special cards, etc, etc.

It's interesting - The Mac, and its operating system, have been changing too much and too fast for information about the changes to percolate through society.

SO many people haven't gotten the message that OS X is a Unix, that it's TCP oriented rather than Appletalk, that it's far more standards oriented, etc.

It's kinda sad that "too fast" in this case means "five years".

Perhaps that shows that Apple could be doing more technically-oriented marketing, rather than aesthetic marketing.

 

[MagnusDredd]

It's interesting - The Mac, and its operating system, have been changing too much and too fast for information about the changes to percolate through society.

SO many people haven't gotten the message that OS X is a Unix, that it's TCP oriented rather than Appletalk, that it's far more standards oriented, etc.

It's kinda sad that "too fast" in this case means "five years".

Perhaps that shows that Apple could be doing more technically-oriented marketing, rather than aesthetic marketing.

I really don't think it'd make a difference. I study processor architecture in my spare time, build Wintendos for myself and others, own machines using 6 different ISAs (68k, PPC, x86, Sparc, MIPS, and PA-Risc), do networking on the side from time to time, administer servers running Win2k, WinXP, Slackware Linux, and OSX server, etc. People who work with me, and are around me are aware of at least some of this.

These same people spout the same misinformation I have corrected on numerous times.

My approach has changed somewhat. When I hear the slow comment, my reply is "Oh, so that's why Virginia Tech was able to build the third fastest cluster in the world using Macs for a fraction of the cost of much slower ones using Xeons. But then again what do scientists know."
Expensive: "Oh yeah, I'm glad I'm not so poor that I can't spend $500 on a computer"
etc.

Perhaps making them feel dumb is the wrong approach but it's the only one that seems to work.

People are strange. They decide something is one way and there is no reasoning with them. It's like they are convinced that things never change... Also most people prefer not to think. *shrug* How boring....

Note: I don't want to pay for a cell chip that I won't use... That would just make the machines that much more expensive. If graphics Pros wished to fork out a couple of thousand for an add-in card with limited usefulness they should do so.

 

[~loserman~]

My approach has changed somewhat. When I hear the slow comment, my reply is "Oh, so that's why Virginia Tech was able to build the third fastest cluster in the world using Macs for a fraction of the cost of much slower ones using Xeons. But then again what do scientists know."

I wouldn't use V-Tech as an example in proving your case.... So far to date System X has not proved to be useful, or stable, and is unsuitable for running production code. Here almost 2 years into the project they have admitted it was fine for a proof of concept machine but it just hasn't lived up to expectations. Currently the system is going through another overhaul and they will be replacing their Infiniband gear and possibly even the OS.

 

[MagnusDredd]

I wouldn't use V-Tech as an example in proving your case.... So far to date System X has not proved to be useful, or stable, and is unsuitable for running production code. Here almost 2 years into the project they have admitted it was fine for a proof of concept machine but it just hasn't lived up to expectations. Currently the system is going through another overhaul and they will be replacing their Infiniband gear and possibly even the OS.

Good to know, thanks... I had not been following it (System X is the final name?) for some time... However the speed issue does hold true, especially in Biotech. The problem is the lack of optimized code.