Slashdot Mirror
Why Apple Picked Intel Over AMD
An Anonymous Reader writes "Macworld has a piece looking at why Apple chose Intel chips over AMD's offerings when it decided to move away from IBM." From the article: "The reason, industry analysts say, is that Jobs has a clear goal in mind: innovative designs. And such designs require the lowest-voltage chips, which IBM and Freescale were not going to make with the PowerPC chip core--and which AMD has not yet perfected 'This is a practical, pragmatic Steve Jobs decision,' says Shane Rau, Program Manager, PC Semiconductors for market research firm IDC. Intel serves up the most complete line of low-power chips for mobile and small form factor computers, and a good-looking future roadmap for it. Also, Intel's mammoth production capacity erases any supply worries. "
With all the talk of voltage and mobility, there doesn't seem to be any mention of the impact, if any, on the bottom-line cost and price factor, which is of obvious importance to both Apple and consumers. Interesting that this comes in just a day or two after the story about Intel chips costing $40 to make.
Andrew Lenahan http://www.starblind.com/
"We're sold out on chip sets," Bryant said during a conference call to discuss Intel's third-quarter financial update. "I think chip sets [will] remain tight into the fourth quarter."
Er, this sure seems like a "supply worry" to me!
\burt
There is no such thing as bad weather - only inappropriate clothing.
I'm really looking forward to OSX on Intel and the sooner the better as far as I'm concerned.
I have a theory as to why Apple aren't coming out with them until sometime next year - I believe they actually want to come out with new machines at the same time as Vista is released. Why?:
1) Microsoft is going to spend (pinkie to mouth) 100 hundred billion dollars on promoting Vista. That's going to make a lot of noise, which Apple can cheaply ride on the back of. Imagine, loads of mainstream publications will cover Vista, and if Apple launches at the same time they'll surely do comparisons.
2) It will be switching time for everyone - current Windows users will be thinking - should I move to Vista? If there is another viable option visible at the same time, then they might consider that too.
3) Steve Jobs may be confident that the next generation of OSX will beat Vista in comparison reviews - hell, the current version (Tiger) has a lot of the features Vista is supposed to have already.
Anyway, that is my theory, which belongs to me and is mine.
It's still significant enough to warrant a discount, which is something AMD can't really bring to the table. Their Q3 2005 results are impressive -- 1.18 million Macs shipped for the *quarter* (35% growth) -- 687,000 desktops; 495,000 portables. and 6.155 million iPods shipped for quarter (616% growth). Now I'm no industry expert, but 1.18 million chips is enough to warrant a discount, hell, lots of 1000 are enough to knock a few bucks off per chip for chains and resellers.
Now imagine Apple keeping this phenomenal growth up, and you get the idea that cost-savings is going to be a big deal at this size.
I wonder if Apple has a preferential deal to get the new, fast parts first. When a new process is being ramped up, there is an initial period where they can make some processors, but not a lot.
Apple being a relatively small consumer of Intel parts could be quite happy with this small volume of fast parts and put out machines that trump the wintel vendor's clock rates.
It is a lesson that Apple learned back in the dark days of Mac clones. Since Apple only refreshes a Mac design a couple times a year people know when it is coming and will hold off for the newer version. When that version comes out there is a big demand spike. To avoid long backorders Apple has to have enough processors in hand to cover the initial orders and enough capacity to keep up with the flow after that. The clone vendors, being a tiny fraction of the Mac market could introduce models with the faster processors as soon as they became available in limited quantities. The double nasty effect was that the clone vendors got the reputation for faster machines since they could bring theirs to market faster and they delayed Apple's ability to get the new xxMhz 68030 to market because instead of stockpiling chips for Apple, Motorola would be selling them to the cloners.
I don't believe the article.
I think it came down to money- in some fashion intel offered them a better deal. I have intel and amd computers and amd has a dramatically better cost/performance ratio. I bet that there is no hard technical reason why Mac couldn't have run on both- if they are going to be Intel only it is for political/financial reasons instead of technical ones.
She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
If TFA is correct, Apple is planning to use the low-power Chips promised for next year, rather than the AMD64 which are pretty good right now. Which is fine if Intel can deliver, but I would not like to bet the company on it. If I was in Steve Jobs' shoes, I might do it the other way round:
Use AMD64 now, switch to Intel later if they keep their promises.
C - the footgun of programming languages
And that is, platform and chipsets.
Yes, Intel has more offerings and better roadmap, and volume discounts, and programmers, and prestige....
But this particular analysis is not mentioning the fact that Intel can give you a system, head to toe. That will allow Apple to move the R&D cost of mobo desing to something else, like SW engineering, or industrial design.... go figure...
Now, if I put on my aluminum-foil-thinking-cap, I can think of the following arrangement:
Intel debuts a new and improved processor/chipset combo, and gives it to Apple with, let's say, six months advantage over everyone else, as beta testers.... If there are no bugs in the combo, all is nice and dandy. If there are bugs in the combo, Intel correts them in the silicon, for all the PC bunch to use, and Apple, having more control over the platform than anyone else in the indutry, corrects the errors via a BIOS/OS patch, intead of a more costly recall.... Match made in heaven! Apple gets a six months edge, Intel gets a HUGE and cheap field trial of new silicon!
Just my two cents anyway....
In the end, there was not just ONE magic reason, but a host little thing that made Apple choose Intel over AMD, Transmeta, VIA/Cenatur and all the others out there...
*** Suerte a todos y Feliz dia!
I'd tend to agree that it is more important that we look for new uses for this new technology from Apple, rather than dwell on why they made the switch. One benefit is in the gaming world. Games that use optimized x86 assembly could very possibly be ported to Mac OS X far easier. After all, a multimedia OS like Mac OS X would gain greatly from high-end, 3D games.
Cyric Zndovzny at your service.
Among the market Apple hopes to secure, I suspect the name Intel is better known & respected than AMD.
We geeks know that AMD has some good stuff, but I'm sure we can all remember when AMD provided chips like the K6/2 which while technically sound (100mhz front-side bus before the Pentium-2 became common, right?), tended to be sold cheap and built into PCs which also used cheap chipsets and reliability suffered as a result.
Back in the day, the P2 and early P3 and the K6/2 and K6/3 were only really differentiated by the quality of their chipsets. A lot of the people Apple wants to woo may have suffered the effects of cheap AMD-based PCs. Intel's late-90s/2000-era chipsets were pretty solid and due to better build quality and drivers tended to run Windows somewhat better.
I wouldnt be surprised if that still affects the market today. Technology moves on, consumers are more static.
Anyway they can always get established on x86 then build an Athlon64 xServe......
I have been a user for about 10 years. This ends Feb 2014. The site's been ruined. I'm off. Dice, FU
Once again, RTFA... it's not about how much power AMD's Athlon64 FX consumes compared to the Pentium XE... The whole article was about low power low voltage chips like the Pentium M. The whole point of the article was that Intel has on the table low power dual core Yonah processors for early next year, while AMD has not disclosed anything about that. Sure AMD has Turion, but that's a single core chip, and it's not fair to compare that to Yonah.
If you payed attention during the Intel Developer Forum, you'll realize that 5x "performance per watt" was compared to Banias, the first Pentium M... which DOES NOT SUCK at performance/watt now.
Intel's Pentium M chips don't suck today... so you're mistaken. The whole article was basically about Pentium M, not Pentium 4 netburst, you buffoon. And moreover, the whole point was future processors, not processors today, so your point about how Intel sucks today is stupid.
Other Advantages would be that it permits Apple to
- ride the tide of CPU speed (no more "megahertz myth").
- pass on processors (try telling IBM that you aren't interested in their minimal speed bump when you are their only client)
- use PC graphics cards without modification
- diversify their product line (if you haven't noticed, the dual G5 is nearly on par with the top of the line Intel... but the middle and lower end systems from Apple aren't even in the ballpark)
As a Mac user, it's a bit hard to swallow that I'm going to have an "Intel Inside" but there are simply too many advantages to overlook. Intel seems very interested in having their processors in everything from handheld devices to super computers... IBM does as well, but do they have the resources?
Stephan
If AMD comes out with a better chip in terms of power usage, Apple can switch anytime.
Unless Apple has an exclusive, multi-year contract with Intel. Or unless Apple is also using Intel chipsets for non-processor HW(switching to AMD ticks Intel off, so they raise the prices for the rest of the chipset, not to mention taking away the advertising kick-backs).
That's a choice they didn't have before with the PPC architecture.
Because IBM and Motorola/Freescale are the same company?
Sorry to be a pedant, but the quote said "voltage" and you went off on "power".
The two are not analogous. Running with a lower voltage (at the same frequency) is based on the properties of the transistors and reflects a more advanced fabrication technique.
I don't know (and am too lazy to check) if the claims in that quote about voltage are true, but if they are then that means Intel has more room for growth in the future (which is probably more important to Apple than what is going on right now).
The article refers to the lowest voltage chips.. This, of course, is one factor that contributes to the real issue: low wattage. The system needs to consume less power and generate less heat.
They also claim that Freescale (former Motorola chip division) cannot achieve these low power levels. I'm not sure where they get this impression from. The PowerPC has always been a low power processor. They are most commonly used in embedded devices, like routers and switches. They keep ratcheting up performance, while trying to keep it under 10Watts.
While the PowerPC's from Freescale won't be at GHz par with the Intel P4's. They aren't far behind the more comparable Pentium M's in clock speed.
IBM, on the other hand, makes CPUs primarily for their workstations. So, their power usage has always been much closer to Intel's..
You don't think Apple had access to the new 970s? I seriously doubt that Apple would go through such a wholesale change in technology without running a benchmark or two. It's a good bet that the new 970s don't perform as well as the new Pentiums clock-for-clock, or else Apple would have stuck with PowerPC.
RTFA. This isn't just about desktops. In mobile performance, Pentium M mops the floor with AMD's mobile Athlons.
Yes, yes, we get it. Today's Pentium desktop chips are hot, power-hungry underperformers. Good thing Apple isn't using today's Pentium desktop chips. (Developer Preview loaners excepted, of course.)
Show me an example of true revisionist history, and you may have a point. But the people you derisively refer to as "Apple zealots" are anticipating the new, lower-power Pentiums that Intel announced at IDF just as much as Windows and Linux users.
This sig intentionally left blank.
"The whole point of the article was that Intel has on the table low power dual core Yonah processors for early next year, while AMD has not disclosed anything about that. Sure AMD has Turion, but that's a single core chip, and it's not fair to compare that to Yonah."
Your right it is not fair to compare a chip that everyone can buy now with one that doesn't even exist yet.
The sad truth is this is beginning of the end of apple as a computer hardware company. They are going to use use Intel motherboards in pretty Apple cases. Pretty much the same thing that Dell does now.
I just hope they keep OS/X.
I find it very sad that Apple is going backwards to the nasty x86.
See my blog http://ilovecookes.blogspot.com/ for light hearted technical information.
Some good points. However, switching chipsets is much less work than switching instruction sets. Didn't some people already have hacked versions of the new x86 OSX running on AMD hardware?
Also, IBM and Motorola obviously didn't find Apple worth fighting over. With x86, even if Intel and AMD aren't competing for the sake of Apple, they'll compete for the sake of the very large x86 market, and Apple will reap the rewards.
The kickbacks, exclusivity agreements, and advertising dollars are another story. But those issues could arise no matter what architecture you're using.
This will probably get lost because it's so deep in the comments, but the reason isn't technical, it's personal.
Apple was unhappy about the direct attacks AMD was making against Apple on the DAW (Digital Audio Workstation) front. Look at all the inroads AMD is making into the music and video business, and some of the negative comments that were made toward Apple. It's not hard to see why they wouldn't get in bed with AMD.
Pretty much. It would perform a little better, particularly on things like MySQL. But for most people the increased performance would be too minor to notice.
And it would have been considerably later to market, since a lot of underlying infrastructure stuff would have had to be ported. That's one reason they didn't do it.
The other reason, of course, being the licensing.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Friends don't let friends enable ecmascript.
Intel's mammoth production capacity erases any supply worries.
Um, no.
Intel has been a constraint on supply to customers in the past, and will be again, because they're not clairvoyant, and maintaining enough capacity to handle 100% of the distribution of order-rate excursions is wasting money (for those who slept through Technology Policy of the Firm: it's like building an 80,000 seat stadium for a basketball team; sure, once every 30 years you'll fill it, but the rest of the time, you're eating your hat).
It may have mammoth production capacity (ever try to keep a mammoth down to class-1 cleanroom standards?) but that capacity is not monolithic nor is it readily fungible. It takes years to do some kinds of process changes, and most chip designs are tuned to a single process and could not be simply adapted to be fabricated on another process.
What this means, if Jobs is any kind of mogul with any sense of supply management, is that Intel will have to build capacity tailored Apple's needs.
Which is good++ for Intel, because their real business is building and filling fab lines; designing and marketing chips is a cost to them.
Games that use optimized x86 assembly
Hand-optimized asm went the way of the dinosaur years ago. Any games *that* old would run fine in emulation anyway. The biggest obstacle to porting games to the Mac is a C++ API - DirectX.
Lost: Sig, white with black letters. No collar. Reward if found!
Look, 10 years ago in '95, Jobs was behind the campaign that promised the new PPC machines were 100x faster than Intel. Remember? He compared SpecINT running a 486 to the latest PPC (or was it a 486 compiled binary, I forget.) Point is: Jobs knows the marketing game, and for him to work with a company he tried to beat in the numbers game means he can TRULY see through the bullshit that other makers ... *dell* **cough cough** cannot.
I know y'all hate Intel, but maybe, just maybe, they got something right and Jobs can smell it.
https://www.accountkiller.com/removal-requested
The OS X kernel, where this DRM would be put in, is open source. It is not inconceivable to easily get around that.
And programers. I've heard there are at least 300-400 people working at Intel on intergration with Apple at the hardware and software level, including programing tools. AMD doesn't have the resources to devote to a niche player like Apple in that regard.
It would be nice if someone who knew more than me would comment on how the xscale processors may have figured into Apple's decision.
I think Cringely may have brought it up a few weeks back.
Davo -- Free speech, free software, AND free beer.
In this case, AMD would have been the "don't be evil" warm and fuzzy choice (see AMD-v-Intel suit). Transmeta would have been the cool-tech choice. Picking Intel was pure cold business rationality.
Jobs doesn't bend other people's reality so much as exercises his power to mould new realities. This is evident in his string of lucrative industry firsts.
(Malone's Infinite Loop is a fairly balanced account of Jobs, rich in background detail, neither hagiography nor a total hatchet-job.)
you had me at #!
I always wondered why Mac decided to change from PPC to x86 and not x86_64? You would think if there going to change they would go with a 64bit arch. I know Intel has 64bit desktop CPUs and the mobile 64bit CPUs should be out soon.