AMD hammer switch
<a href="http://www.amdzone.com/articleview.cfm?ArticleID=1212" target="_blank">http://www.amdzone.com/articleview.cfm?ArticleID=1212</a>
Maybe this has already been posted in another thread but it seems like Apple is going to have ample choice ...
Maybe this has already been posted in another thread but it seems like Apple is going to have ample choice ...
Comments
<strong>I'd assume that the PowerPC 64 is the 970. The reporter doesn't realise the magnitude of this point: what happens to all the old OS9 apps and third party OSX apps?.</strong><hr></blockquote>
Old OS 9 applications of course would not run on any x86 implementation of OS X. Third party OS X applications would be recompiled after any endian issues were identified and resolved (which for most programs is quite simple to do).
I think your life has been good enough untill now. Let's see some NDA busting postage in here: ever seen a working OSX on AMD (main CPU) prototype?
That can't just be coincidence. (sigh)
...not for a while. Certainly not while the X migration is still fragile, and certainly not while THE 80% OF MAC USERS are still in 9. They need an upgrade path, which is Classic.
Moki is great, but let's not start reading his mind and divining from bird entrails and sh!t like that.
And yes--the author of that article knows very little about Macs.
Apple COULD move to Hammer, but it would be easier to go with the (PowerPC 970) flow.
If Moki breaks his NDA, he won't ever recieve another one to sign. And then he will not be able to "intelligently speculate" ever again. Is that what you want, is it? I know, the answer is yes.
"Apple hasn't announced anything so that means they are still considering" (paraphrase). HUH? Since when did (the Steve J.) Apple ever pre-announce future CPU plans???
Barto
I wouldn't rule it out. An x86 transition might be tricky now after developers have 'just' got round to 'X'ing everything...well 'almost' everything. However...what about in two years time? Nothing in the computer industry is cast in stone.
Who'd a thought of a digital appliance for x86?
Now, come on, who'd a funk of iPod on Dell's webstore? Think the impossible (Apple's next ad' slogan?)
There are x86 whispers now. There may be a grain of truth in these chinese whispers.
However, Apple themselves are saying PPC is what they're doing right now.
Recent interview has a quasi-quavering conviction about the PPC. An almost 'temporary' tone.
Apple seem to be making it perfectly clear that once the transition to 'X' is complete then OS is quite portable. And when Apple themselves start talking like that then people should take note!
What does that mean? It may mean that if Moto or IBM start dickin' them around in the future...they'll move to someone who wants their business!
It may mean Apple will offer the 'X' on more than one processor.
'X' on x86 Hammer? A gut part of me still can't rule it out. Especially if, somehow, it's part of 'diversifying' strategy that is concurrent with PPC.
There's a large Linux crowd on x86. Maybe Apple wants a slice of the cake with 'X' on Hammer x86? They only have to sell alot of boxes with a 'temporary' solution to get their foot in the door. Things change in I.T very quickly.
Take Apple's Xserve. From 0% to right behind Sun M S. Apple now has a slice of the server market. Who'da ever funk it..?
Lemon Bon Bon
Question: wouldn't Apple have to scrap the Carbon API's for a x86 change over? Not everyone is getting on the Cocoa bandwagon it seems in the developer arena. Wouldn't that make a switch even more painful?
Anyhow, for all I know the PowerPC architecture is really nice and shouldn't have problems keeping up with x86 once that elusive 970 is out. Having said that, I run a 700mhz Athlon at home because I couldn't afford a Mac back when I was shopping for a home computer. And it runs fine so unless something breaks for good I'm going to keep using it. After that, I'll get a Mac.
<strong>Hey, long time lurker, first time poster.
Question: wouldn't Apple have to scrap the Carbon API's for a x86 change over? Not everyone is getting on the Cocoa bandwagon it seems in the developer arena. Wouldn't that make a switch even more painful?</strong><hr></blockquote>
The Carbon API would survive more or less unscathed. In fact, abstracting the libraries away from one CPU architecture would be a great reason to dump some of the cruftier bits of Carbon included for legacy compatibility, and streamline it into its emerging role as the low-level system API.
OS 9 would fail to run, and Classic apps would have to run under emulation to run at all. But Apple's done CPU emulators before.
I still don't really see the point of doing this, except perhaps to allow Apple to release a big, hot 3D powerhouse to tide them over until the 970 shows up (and maybe beyond - a lineup that went from G4 and Athlon to 970 and Opteron probably wouldn't piss too many people off. But there is no possibility that the whole line will go x86. The G4 is a much better choice for notebooks and iMacs (and soon iBooks).
Video card compatibility and the like should not be a problem: There are a lot of reasons why PC hardware can be so finicky, and the CPU is rarely one of them.
<strong>You know, one thing that scares me is that I've never seen Moki criticize, or even have any problem with, the notion of Apple moving to AMD.
That can't just be coincidence. (sigh)</strong><hr></blockquote>
i think this is because moki hasn't to be afraid about something which could hurt him or ambrosia-sw -> he's a very good programmer and therefore wouldn't have a problem with coding for os x on x86... BUT i don't think that apple will make this step. just if they can't sell any machines themselves and i don't see this happen anytime soon (or late)
The PPC 970 is very definitely a prime candidate for the next generation of Macs, but the one variable that no one seems to be able to hammer down (no pun intended) is the price range. If this chip ends up costing $1000 or more, that's going to put Apple in a real bind. They need to either keep the prices in check and boost performance in a big way, or even lower them a little more and boost performance.
Increasing the price of all Power Macs by $500 or more isn't going to help their bottom line - no matter how well the machines perform. Especially in a lousy economy (if we're still in one a year from now - who knows with all this war on terror crap going on).
This much is certain: if Apple ever DOES go to an AMD solution, they will still be designing the accompanying hardware - and they'll still be the sole producer of the machines themselves. One thing Apple MUST have in order to pull off an OS X to x86 switch, is control over the hardware environment. There has to be standardization in terms of the video and audio, in terms of the bus topology and other critical factors. Otherwise, the famed reliability and crash-proofness of OS X is going to take a big hit.
They will *need* every conceivable selling point for getting x86 hardware types (not to mention... US) to move to OS X, and one of the very biggest is the stability / crash-proof thing. No Windows OS comes close to matching OS X's uptime. If they lose that to any meaningful degree, they lose their leverage IMO. Early on, there is no way that software and peripheral choices will be a selling point because Windows platforms will clearly have more choices for the first year or so.
It likely won't be price either. I still think Apple's machines - AMD or not - will be pricier than typical Dell PCs. That's not to say the prices won't go down - they will. Significantly. But they'll likely still be higher than what Dell can afford to sell their stuff at today, for example. [Apple must build the hardware, but they must also recoup their R&D expenditures to some degree.]
Bear in mind also: Apple giving the appearance of going to AMD serves [right now] to do only one thing - light a fire under the ass of Motorola and IBM. If they leave PPC, those companies lose some big business. Even if they butt heads with Apple, they still want Apple's money. Until now they had Steve by the nuts, but it appears that role might be reversed in the next 12 months depending on new developments from Apple and the % of users who have moved to X by next summer.
I've always thought OS X on x86 was one of the dumbest ideas I'd ever heard. Mostly because developers like Adobe would be forced to port their applications [to a new platform] *again* - even after all the migraines of moving to OS X at the behest of Steve. Do they really want to spend on the [human resources and money] required to do this, all for a market that technically doesn't exist yet - [or go through another "We STILL don't have Photoshop for (OS X on Hammer)" from Steve??]
It all hinges on the developers. If Adobe says "no, we're not going to build apps for OS X on Hammer until you have so many million users in the design market running this hardware" - Apple can't do it. Period. It's [would be a] catch-22. Developers would't do it until there are a critical mass of users on the new platform - users wouldn't go to the new platform until developers were on board.
If someone can explain how Apple could manage to solve that problem without forcing developers to take a financial hit, I'm all ears. Even so, it seems Apple now might be forced into this, depending on the price of the 970. Things they are a changin'.
(edited for additional thoughts / grammar)
[ 11-01-2002: Message edited by: Moogs ]</p>
I put off building my new PC for 6 months because I wanted to put an Opteron in it. Then, AMD announced that the chip would be delayed. Then, at the Microprocessor Forum they stated that the Opteron was initially being "de-emphasized" for the desktop in favor of server applications. I'm not really certain what's going on with the chip, but it's not really encouraging that there continue to be delays. All other issues (recompilation, etc.) aside, I think that Apple would be better off with IBM for now.
Maybe AMD shouldn't have cozied up to Motorola. You sleep with dogs and you get fleas, you know!
Now software... Adobe, Microsoft, Macromedia, Quark, Intuit... You name it. Nearly ALL Mac software developers also develop for Wintel. Imagine if these companies could maintain a single code-base to support all platforms. Theoretically, the hardware is the same. I know there's much more to software development than targeting the hardware, but that's just one less thing to worry about.
Suppose it's possible to compile apps from the OSX dev tools to Windows and all other *NIXes. To some degree that functionality already exists, does it not? Perhaps Apple takes the 'digital hub' concept to software development. What if OSX was the best dev environment for creating Windows apps?
Now... Add the oft-rumored Windows compatibility layer to OSX. So we get 'Classic' through emulation, OSX native, and Windows much like we now get Classic. Suddenly you have your choice of OS on your box WITHOUT partitioning. Now your Mac CAN run the latest and greatest games - on time and at speed. That alone would open the Mac platform to all of my friends who haven't yet switched.
Sure, Apple would probably have to license from Microsoft - but Apple sells Boxes, Microsoft sells software. Nothing changes, except they both gain new markets. MS has more access to Apple's 5%, while Apple gains admittance to Window's 95%. Good trade. Does MS care? Not if they get their cut.
Moving away from the Desktop, we enter into Workstations and Servers. With this Apple uses Power4/5s and PowerPC 97x.
Take these multi-core chips, stick a bunch in a box with lots of fans, and utterly dominate high-end graphics/science/CAD/editing/etc. Get high-end graphics card makers to supply video cards, ensure Maya, Photoshop, FCP, Shake, AVID, Lightwave, et al run amazingly fast and reliable in these boxes, and you own the graphics market. (Just make sure Apple kills cheap Linux boxes on the performance front. - Unless they buy Apple Workstations, then load Linux on them. That's fine, too.)
With servers, continue to produce fast, reliable servers/render-farms/clusters for the Workstation & Desktop customers AND license OSX Server to Big Blue to run in their Low to Mid-Range servers. Or, better yet, BUILD servers for IBM. They seem more interested in services anyway. Same chips, same tech, same market. Why not?
One last thing... I know hobbyists like to built their own kit. I do. Only ship iApps with systems, and sell the OS alone. Sales of iApps might make up the difference in lost hardware margin. Perhaps they also opt to sell there own Motherboards, or better yet an assortment of first-rate computer enclosures. No? Just a thought.
None of this is likely to happen, given. Some of it isn't even possible. It does, however, illustrate that Apple COULD use Intel/AMD chips without going out of business or losing its identity. IBM successfully sells multiple OSes on multiple hardware platforms. Why couldn't Apple?
A 64-bit update to the Athlon core with onboard memory and hypertransport.
Same idea as the near-mythical G4.5, take a bandwidth starved core and give it more bandwidth.
Except the G4.5 doesn't need to be 64-bit as the PowerPC/VMX ISA doesn't suck.
AMD took the easy way out, and they may regret it.
The Pentium4 may only dispatch 1 instruction/cycle, but its effeciency (sustained processing closer to peak) and very high clock speeds beat the Athlon.
Increasing bandwidth may help AMD, but without better manufacturing and an new core, the Hammer might be a financial black hole.
Barto
[ 11-02-2002: Message edited by: Barto ]</p>
Fast, yes. As fast as the PowerPC 970 will be, no. x86 is more expensive than the G4.
"Apple already uses standard PC components throughout its architecture. DDR 2700, PC-100/133, PCI, AGP, USB, Ethernet, IDE, etc. The only main core NON PC component is the CPU."
Standard ≠ PC. Apple has every reason to use standards to make upgrades and devices work on both Macs and PCs, but if they are going to use an radical OS, they may as well use the best CPU. Which Apple believes is the PowerPC (in a long-term vision).
"If the Marklar rumors are true, Apple even has an x86-capable OS."
They are true. Replace the PPC Mach kernel with a x86 Mach kernel in the OS X CD, and boot in a PC system. It boots until "your CPU is not supported". Apple DVD Player.app has a line of code; "Disable Debug PIII PIV ATHLON". There is big rumors inside Apple of the implications of Marklar.
"Now software... Adobe, Microsoft, Macromedia, Quark, Intuit... You name it. Nearly ALL Mac software developers also develop for Wintel. Imagine if these companies could maintain a single code-base to support all platforms. Theoretically, the hardware is the same. I know there's much more to software development than targeting the hardware, but that's just one less thing to worry about."
I want applications with a Mac OS X interface. If I want Windows apps, I'll run Windows. If I have taste and want good interfaces in applications, I'll buy Mac OS X. What your proposing eliminates the key advantage Macs have over Wintel. ITS ABOUT THE SOFTWARE, AND MAC SOFTWARE IS BETTER.
"Suppose it's possible to compile apps from the OSX dev tools to Windows and all other *NIXes. To some degree that functionality already exists, does it not? Perhaps Apple takes the 'digital hub' concept to software development. What if OSX was the best dev environment for creating Windows apps?"
I've thought of that myself... It would be a possible direction for Apple. Instead of DevTools being necessary for Mac OS X, DevTools 2 would be THE development environment for Wintel, Linux and Mac OS X. Maybe an Apple HIDE (Hardware Integrated Development Environment) which accepts x86 and PowerPC CPU cards, but moving the entire Mac platform across to x86 has been discussed to death. It ain't happening anytime soon.
"Sure, Apple would probably have to license from Microsoft - but Apple sells Boxes, Microsoft sells software. Nothing changes, except they both gain new markets. MS has more access to Apple's 5%, while Apple gains admittance to Window's 95%. Good trade. Does MS care? Not if they get their cut."
Apple does what IBM tried to to. Sell boxes with a Windows compatible OS. That failed. People buy Windows because they are sheep. Sheep won't buy Mac OS X if it runs Wintel apps, because it is still different. The only way Apple will, without becoming a PC-clone maker, ever get marketshare over 10% is if Microsoft collapses.
"Take these multi-core chips, stick a bunch in a box with lots of fans, and utterly dominate high-end graphics/science/CAD/editing/etc. Get high-end graphics card makers to supply video cards, ensure Maya, Photoshop, FCP, Shake, AVID, Lightwave, et al run amazingly fast and reliable in these boxes, and you own the graphics market. (Just make sure Apple kills cheap Linux boxes on the performance front. - Unless they buy Apple Workstations, then load Linux on them. That's fine, too.)
With servers, continue to produce fast, reliable servers/render-farms/clusters for the Workstation & Desktop customers AND license OSX Server to Big Blue to run in their Low to Mid-Range servers. Or, better yet, BUILD servers for IBM. They seem more interested in services anyway. Same chips, same tech, same market. Why not?"
A more practical approach would be to license Mac OS X to the entire IBM range of workstations. IBM does that kind of hardware better than Apple probably could.
"One last thing... I know hobbyists like to built their own kit. I do. Only ship iApps with systems, and sell the OS alone. Sales of iApps might make up the difference in lost hardware margin. Perhaps they also opt to sell there own Motherboards, or better yet an assortment of first-rate computer enclosures. No? Just a thought."
Suppose Apple introduced a low-key (only known about in the enthusiast world) motherboard and OEMs are prevented from using them. It would have to be totally unsupported motherboard, but even then a large chunk of mac-dom would probably start to use them.
Maybe an expensive developer motherboard could be sold, costing $500 or so. Only the true enthusiast would shell out that much. It would take standard PowerPC CPU cards.
"None of this is likely to happen, given. Some of it isn't even possible. It does, however, illustrate that Apple COULD use Intel/AMD chips without going out of business or losing its identity. IBM successfully sells multiple OSes on multiple hardware platforms. Why couldn't Apple?"
Because IBM sells crap Wintel PCs, decent Wintel laptops, Linux/i386 entry-level servers and workstations and PowerPC high-end servers and workstations.
They are all in totally different markets. The only market Apple isn't in is the very-high end. Why not just sell IBM a Mac OS X Server license and have that base covered.
Barto