Apple, why Intel, not AMD?

I think competitive pressure is a good thing which benefits us users, be it NLE, CPU or OS. Hopefully, what AMD offers now will spur Intel to do even better by the time Intel chips appear in Apple computers. I hope.

But this is one version of the situation as it exists now:

[www.digitalvideoediting.com]

Oh boy, more pointless rendering/cpu speed/benchmark bull puckey tests....

My bottom line has been and always will be "Can I do what I need to do creatively without interference from the OS?" If the answer is yes, than I will create my vision on whatever frickin' platform does it the best...and in this case, that's a Mac and has been for years (and yes, I have run Avid, AE, Boris Red, Inferno et. al. on PC workstations - the simple act of FINDING SOMETHING on the system drove me away from PC's and keeps me away)...

end of story.

<< Apple, why Intel, not AMD?>>

I'm pretty sure it boils down to five things:

1) Intel is able to offer a better supply of chips than AMD, which occasionally has problems.

2) Intel's roadmap could have been more impressive than AMD's.

3) Intel does more than AMD. Maybe they were able to offer "whole package" motherboard (aka logic board) design that AMD wasn't able to offer (does AMD even make mobos? I don't know.)

4) Intel has been a bit pissed at M$ lately. Maybe this spurred them to court Apple harder.

5) Jobs already has a relationship with Intel from his Pixar server farms. Intel (non-CPU) chips are also already being used in Xserve RAIDs.

The history has been that Intel and AMD change the "speed king" title back and forth over time. Who knows who will be faster when Intel Macs are finally released. And when the switch to x86 is finally complete, there will probably be very little stopping them going to AMD if that proves more advantageous at that time. IMHO, Apple made the right choice of the two.



- Justin Barham -

What is hardware speed worth?

If it allows me to get 5 minutes more worth of productive work done during the day, and my day rate is $150/hr then it takes somewhere around 112 work days pay off a $1400 premium on a machine.

I've had a little time to reflect on Job's announcement - and I think Apple is in trouble if they indeed take the full two years to switch. At this point we need a point update for tiger that runs on an intel or AMD within the next few months.

The low end macs are still going to be worth getting for Mom, with under $200 in software - they will be fine. That's the good news. The bad is there's no way I see the profit potential in buying a high end Apple any more. If you've got a dual g5 now there's not much of a way to make the $ work out before the machine is dead ended in 2 years. If it were just the hardware I wouldn't care. It's the software and investment in work flow that combines to dead end Apple until they offer a proven road.

Apple needs to demonstrate that their promises of a painless transition are going to be real by significantly beating all posted schedules for the switch. Once Apple demonstrate that OS X works on an Intel chip - the future looks bright for Apple.

Ian

Apparently OSX has always worked on an Intel chip. During the demo it was stated that all OSX releases have been compatible with the Intel all along.

Which means that for the last five or so years they have been known they need to be building software for Intels alongside PPC compats. Which, surely, means that FCP would already exist in an intel compatible form. Somewhere.

This is from a conversation I was involved in on another board but I feel it is the most relevant aspect of the Apple switch to us Pro Ap users running PPC Macs over the next two years.

?With the X86 Pro boxes coming out last in 2007, programs like FCP which depend on Altivec which wouldn?t port to X86 (please correct me if I?m wrong) run into a development catch 22. FCP is seeing significant respect in the industry but to keep that trend ongoing evolution is necessary. Dumping resources into development of aspects of software that wouldn?t effectively port to the new processor is a dead end. This creates a challenge for Apple and how they handle it will have an effect on Mac Pro Ap users over the next few years.?

Richard
2K-Plus
Atlanta

Apple tend to write accelleration libraries which, currently, have underlying PPC Altivec and SSE3 (I think) Intel code in them.

Indeed, the hard part of vectorising some code is not the implementation in Altivec or SSE3, but the figuring out how to do it as vectors in the first place. I don't see this, therefore as a roadblock to FCP or other pro app being moved over to Intel. It might not have been done yet, but it's not a herculean task.

Graeme



[www.nattress.com] - Plugins for FCP-X

"not a herculean task."

From you - I'll accept that assertion. Not from Jobs.

Still . . . I would be far far more comfortable if I knew of someone with a working 3rd party Intelized version of OS X Apps.

Graeme - where do you see the economics of this change - do you agree that it's a little tough to justify a new G5 and another full suite of pro software now? If it's as easy as it's supposed to be - do you think it's such a good idea to wait two years for the change?

Ian

I dont see any problem with it Ian as long as we believe what Apple is saying. They said they will be supporting BOTH PPC and Intel for years to come, meaning they will make 2 versions of software, that will run native on both chips. Only reason I might go to Intel in 2007 is if that machine was BIG TIME more powerful than my current one and warranted the purchase. Rosetta would take care of my other apps for awhile.

I just bought a dual 2.5 and intend to use it until 2008. Three years is usally all I can get out of these things anyway. However I'd have no problem buying a new G5 in 2006.



Michael Horton
-------------------

The situation will be further complicated by the transitions to HDV/HDTV and true 64 bit computing. And 'the competition" will be a factor especially for potential new FCP users. What will the Avid buyout of Pinnacle bring to the table? (Bench?)

Perhaps the Intel chips will allow FCP to grow outside of the box and begin to take on more characteristics of Silicon Graphics editing techniques.

I think a good mac bought today will still be a good mac in 5 years time. I see no reason to hold off a purchase today. With video editing, the technology is moving still very rapidly, and a good mac might not last as long if you want to stay at the top of the pile - say 3 years, which will more than bridge the gap to the new Intel Macs, and indeed, PPC Macs should still be very supported in 3 years time.

However, the move to PCI-Express will make old macs get much older, much more quickly than any move to Intel from PPC will, especially when FCP becomes GPU enabled with PCI-Express so you can see what you're doing on your broadcast monitor in realtime, not slowtime.

Graeme



[www.nattress.com] - Plugins for FCP-X

I'm leaning toward thinking that buying a hot DP G5 a year from now might be a great move, because the price is bound to go way down due to decreased demand and anticipation of a new product line. It will still be as good a machine as it is today. Even though it has hit a performance ceiling, it's still a pretty respectable ceiling (and will outperform my Powerbook by quite a lot). The nature of my work (at this point, anyway) isn't threatened by not keeping up with the latest and greatest.

Another thought comes to mind. Do you suppose the limited internal HD expandability of G5s is due to the CPU heat problem that Apple just can't seem to solve? Is that what drove the limit, or was it some other peculiar engineering decision?

I'm a little puzzled by all the panic over the idea that Apple is pulling the plug on IBM chip users. I give Apple somewhat more credit than that for market wisdom (and I hope I'm right).

Scott

Scott Taylor wrote:

> Do you suppose the limited
> internal HD expandability of G5s is due to the CPU heat problem
> that Apple just can't seem to solve? Is that what drove the
> limit, or was it some other peculiar engineering decision?

Considering third party companies such as Weibetech can put four more drives in there, I'd say no.

mike



Michael Horton
-------------------

The Wise Words of Ian Graham

"I would be far far more comfortable if I knew of someone with a working 3rd party Intelized version of OS X Apps. "

Ian - in fact Mathematic did the port in a few hours....pretty dramatic demo at Steve Jobs Keynote - which, I again, recommend everyone watch

[www.apple.com]

He gets to the intel stuff about 23:05 into the stream - you can skip the beginning - it's just an ode to Mac's greatness - which most of us know

at the tail end of that - the Mathemetica programers come out and show their magic - running their OS-X program ported to Intel chip - in one day - took the rest of the weekend off - if other programmers have it this easy for a program that complex - this shouldn't be the nightmare many expect. Andy

<<I'm leaning toward thinking that buying a hot DP G5 a year from now might be a great move, because the price is bound to go way down due to decreased demand and anticipation of a new product line. It will still be as good a machine as it is today>>

This tends to my thinking as well. The used market will be especially good when the Intel Macs start selling. And you get Altivec, and still get to have Open Firmware, as opposed to the ugly x86 BIOS. I don't see PPC support dropped for quite some time. What will happen though is that programs that require a great deal of Altivec and G5 tweaking (like FCS) will not get that attention anymore as developers focus on optimizing Intel code. For most apps though, the Universal Binaries will be easy to clcik-and-compile.

I'd be more concerned about buying a laptop computer now, with a big performance leap coming in about a year.

<< the Mathemetica programers come out and show their magic - running their OS-X program ported to Intel chip - in one day - took the rest of the weekend off>>

True, but I also understand that Apple sent a crack team of programmers over there to help the effort.



- Justin Barham -

Incidentally. I linked to the 2nd of two pages of the Charlie White article. To see the first page, scroll to the bottom and click on "Prev Page" at the right.