If the X86 are currently better it's because of the great teams of engineers who develop them, not because the X86 code is inherently better : it's just the contrary. </strong><hr></blockquote>
Well, like I said, where there's a will, there's a way. What gave these engineers the incentive to push the x86 farther than what it was intended? Red Bull? Free GAP clothes? :cool:
Well, like I said, where there's a will, there's a way. What gave these engineers the incentive to push the x86 farther than what it was intended? Red Bull? Free GAP clothes? :cool: </strong><hr></blockquote>
I just say that IBM has the will, Mot did not have it. The PPC 970 is the celeron version of the Power4. The Power5 is already on the way : there will be a future celeron or duron like version of this chip for the mac. This line of chip has a future.
Like all celerons like chips, the 970 chips does not represent a terrific amount of R&D at the exception of the two altivec units wich represant a great amount of R&D. But the job is already done, and less R&D will have to be done in the future developpements of the PPC 970.
In other way the power 4 has required a vast amount of R&D : it was a 4 years project developpement. The power line is a huge locomotive that drag behind him the PPC 970 line.
<strong>I don't even know what a low-end server market is.</strong><hr></blockquote>
But you still argue about it? Wow.
[quote]Originally posted by MacLuv:
<strong>Is that the entry level market that everyone with a clue talks about? Because that includes servers ~ $100,000.</strong><hr></blockquote>
Low end as in cheap, low cost servers that still are servers and not PC boxes running Linux. ~ $100,000.- servers are now low end since they pack quite some performance already. You could probably abstract it to the level that low end servers are 1-4 CPUs. But then again, for people deadling with servers worth $10,000,000 low end might be a 100,000 server.
Which, in turn, IBM does not want to be in vain. Hence this means they have longer-reaching plans for the platform.</strong><hr></blockquote>
I think they have plans for the high performance servers based upon power chips and the low end ones based upon the PPC 970. You will notice is that if Apple was not concerned by the PPC 970 they would not have developped a SIMD unit Altivec compatible. A simple core power 4 without altivec unit would have been simplier to produce and design.
I think that producing PPC 970 is a way to recover a certain amount of R&D they spent on the power line.
I was being facetious. Since you guys want me to specify "x86 ISA" rather than "x86 processor" perhaps you should use the term "entry-level" instead of "low-end". Low-end generally means cheap as in manufacture, not price. Actually I don't care which you use, as long as you know.
[quote]I think that producing PPC 970 is a way to recover a certain amount of R&D they spent on the power line.<hr></blockquote>
The immense R&D is part of the reason that Power4 is so expensive (other factors being low volume compared to desktop processors and the supporting arhcitecture being beefy).
so you're saying that Apple is entering the PC server market?
</strong><hr></blockquote>
No, xype got it right: I'm saying that they're entering the low-end, UNIX-like server market. There are entries here from just about everyone, some running on platforms other than x86 (although x86 is common here), and there is lots of competition from the Windows server market, which has a large low-end presence.
It's basically the sub-$10K market. That's the neatest way to identify it. But if you want to segregate by OS also, then Apple is competing with Linux, BSD, and Solaris/x86 solutions under $10K.
No, xype got it right: I'm saying that they're entering the low-end, UNIX-like server market. There are entries here from just about everyone, some running on platforms other than x86 (although x86 is common here), and there is lots of competition from the Windows server market, which has a large low-end presence.
It's basically the sub-$10K market. That's the neatest way to identify it. But if you want to segregate by OS also, then Apple is competing with Linux, BSD, and Solaris/x86 solutions under $10K.
[Oh, by the way: I suppose "entry level" would be more precise, but "low end" is still accurate. You pay for robust hardware, so the best you'll see sub-$10K are consumer-grade components like the x86, with some redundancy here and there.]
Hmm... Although this thread is going way off anything, really, I have to add the following to the discussion...
Whether Apple is entering 'low end', 'low price', 'UNIX' or 'UNIX-like' - I don't care. They're entering the server market. Like Steve Jobs said, they're entering it humbly.
Why are we discussing IBM's and Sun Microsystems' big iron servers when talking about what Apple should or shouldn't do? I don't see Apple in _that_ market in the near future (the next five years). And if I want a stable and fast web/file/mail/print/whatever-server for 'small' to 'medium', the Xserve is a nice alternative to what's around, if you consider price, usability, stability and power.
I'm sure the PowerPC 970 will be the better option for the Xserve than a Motorola chip, but an X86 chip (define that yourself) isn't a better option, unless you suggest that Apple would go X86 for the other product lines, too. But that's basically out of question for the near future (2003).
What Apple wants, in my opinion, is to build a strong option for businesses to go all Mac from server to desktop to notebook. They have a strong selling point there, but they're not well accepted in that market. This has nothing to do with Intel or AMD processors, but rather with how the company is viewed from a business angle. Apple looks like a consumer company to businesses. Very much like Sony. Xserve has made a good entry into the market (I don't really CARE about exact market share figures, this thread shows how irrelevant those statistics are, because it very much depends on how you look at them.). It's gotten good reviews, it fares well in its price/power segment and people start to look at Apple differently. Maybe in a year or two, Apple will be taking a bite out of Dell for business sales. And business people could care less about what processor is in a machine. ROI is important. TCO is important. You don't see many businesses upgrading processors every other months, only because Intel claims that 'the internet is faster with MMX' or anything like that. But they _do_ see that Microsoft is a one-way street. They _do_ see that Microsoft is starting to behave strangely with licenses. And Microsoft pissing off people is an opportunity for Apple.
"I'm not sure what their deal is, but we should be seeing 1.2GHz to 1.6GHz from the die shrink."<hr></blockquote>
The 64 million dollar question is WHEN. There have been very few rumors regarding this lately.
OFF TOPIC
[quote]Originally posted by MacLuv:
<strong>
The whole point of getting into servers and marketshare, etc, was to talk about Apple's market penetration into the server world.
</strong><hr></blockquote>
My bad. I mistakenly thought your original comments regarding IBM's small market share and slight decline in UNIX servers was in response to positive arguments made for Apple using an IBM cpu, because of IBM's large size and ability to spend billions on research.
After your statement of facts, it seemed the argument changed from IBM's commitment/ability to design/develop/market a competitive cpu to Apple's initial server offerings. Regrettably, this confused me.
<strong>I was being facetious. Since you guys want me to specify "x86 ISA" rather than "x86 processor" </strong><hr></blockquote>
That was just me being pedantic, actually. I don't speak for anybody else here, and nobody else here speaks for me.
All this talk of UNIX vs Linux markets is really a little bizarre, IMO. For starters, Unix isn't one operating system -- there is about one flavour per vendor. They are all more or less compatible, except that they run on different processors. Then there is the FreeBSD crowd, the Linux crowd, and now Mac OS X. All of these are quite similar, but they are all different in various details. A lot of software can just be recompiled across the various flavours, and administration/user knowledge is largely transferable but its not like Windows or Macintosh in that you can buy shrink wrapped software and expect it to run on any Unix-like OS. In the server market that is acceptable (and perhaps even expected), but its hopeless in the consumer desktop market. In the server market, therefore, IMO the operating systems should be lumped into Windows and Unix-like. In the desktop market its really just Windows and Macintosh (with the various versioning issues inherent in each of those). Anybody else is in the geek market and is so individualistic that it doesn't matter.
I'm amazed that there is anybody left who doesn't believe we'll see a 970-based Macintosh from Apple at some point next year. That processor officially exists, it has VMX, it is aimed at the desktop, Steve Jobs himself has said that they like what IBM has lined up, etc.
There might be another G4 revision (0.13 micron is strongly hinted at), but who knows how fast it'll be. Moto's scaling problems could persist and 1.3 GHz might be the limit despite their earlier expectation that they could claw their way up to 1.8 GHz eventually. Sure the 0.18 -> 0.13 process shrink is expected to deliver a 30% performance boost, but that is a guideline not a sure thing. Regardless of what they actually get out of it, the G4 is likely to continue to drive the consumer & notebook lines for a while yet... at least until the 0.09-micron 970 arrives.
IBM has built a 64-bit PowerPC before, remember -- the 620 and 630 were designed and produced around 1997 and used by IBM's AS/400 group IIRC. The POWER3 & POWER4 were both 64-bit as well. Lots of experience and tools to be had there. PowerPC was designed from day 1 to have a 64-bit version that was backward compatible with the 32-bit version. x86 wasn't even designed to have a 16-bit version!!! Its been one hack on top of another one.
The x86 market will find its way despite themselves... they've got enough money to spend their way to a solution. It doesn't take willpower to build an advanced processor, it just takes money. Lots of it. Intel has it, IBM has it (and a better starting position), Motorola doesn't have it, AMD might have it. Who are you going to bet on?
<strong>Like Steve Jobs said, they're entering it humbly.
Why are we discussing IBM's and Sun Microsystems' big iron servers when talking about what Apple should or shouldn't do? I don't see Apple in _that_ market in the near future (the next five years).</strong><hr></blockquote>
To be honest Apple doesn't have the resources to enter any bigger a server Market than what can be achieved with up to 4-way CPU servers (lowe end) because the higher segments demand, like Amorph said, quality that has to be proven and that costs a lot. IBM has a reputation of a few decades they can build upon and that's ecactly the reason people with loads of cash to spend but IBM hardware - because for 10.000.000 you better get something that works - and works the way you want. Which is also the reason SGI stayed afloat, because they went back to do what they know best - powerful computing. And not "imagine a Beowulf cluster of these" wannabe power-computing.
The Rack-mountable approach Apple is taking is perfectly fine because it's a market where they could in theory live off their current customer base and compete with cheap Linux boxes. Plus the R&D cost is not really high so the risk is far lower than going head to head with IBM/SGI/Sun for the real server market. Just like their RAID system is neat for the Apple market, but no way they could ever challenge EMC and Hitachi.
Apple is playing some well tought out moves lately and if it weren't for their bus/cpu cripple they'd be in a really good position.
<strong>The x86 market will find its way despite themselves... they've got enough money to spend their way to a solution. It doesn't take willpower to build an advanced processor, it just takes money.</strong><hr></blockquote>
I will argue that - Intel really doesn't have enough willpower to go for something better, they were really content with simply pushing mhz for a long time. Besides, the goal behind most of x86 hardware designs is low cost cheapcheap and that alone will make the platform stay unappealing for most "serious" computer people.
PowerPC always seemed to me like a platform where the functionality was given higher priority than cost/mhz, which is a good thing. It may have not always been in Apple's best interest (IBM servers, Motorola embedded) but if IBM wants to go the workstation/desktop route, they have a really solid base to build upon and a better starting point than the x86 world. Heck, I doubt Apple will ever have problems because or fubared IRQs or because the motherboard will have conflicts with their graphics card. Unless they switch to AMD.
That was just me being pedantic, actually. I don't speak for anybody else here, and nobody else here speaks for me.
</strong><hr></blockquote>
No, it's fine that you pointed out that perhaps I should be referring to "x86 ISA" because--as I went back to examine if it would have any relevance to my agruments--I learned more about how the instruction set would impact the future. For example, as the x86 ISA "moves forward", there's always going to be a billion dollar industry for x86 legacy support. To an idealist this is absurd, to an entrepreneur this is opportunity.
Regarding "Linux vs Unix" variations, etc... industry analysts have to draw the line somewhere. They can't sit down and nitpick over which side Kirk's hair was parted in episode #37. This doesn't mean I don't agree with what you and others are saying--that there are "grey" areas of distinction--but for business purposes the market has already been segmented. I didn't do it, Gartner's & others have. There are plenty of subs that may be spitting hairs with markets right now but I haven't bothered to check. In that the Xserve is under $10K and everyone is claiming that this is a defined market, I ask for help finding out who has defined that market... I don't have the time right now.
<strong>It's basically the sub-$10K market. </strong><hr></blockquote>
I could see them reaching up to ~ $15,000. But only if the XServe & RAID version work well. (Initial reviews & reports on the XServe appear to be favorable) Some of the cheap blade server approaches would work well with a low heat chip, Apple could do one well, and the bioinformics groups (at the least) would love them. It seems like that the few big problems that are easily vectorizible have a boundless appreciation for computing power. More AV units in a box = good thing.
The part I'm curious about is whether the relationship between IBM and Apple might end up being close enough for IBM to offer something that Mac OS X can run on easily. Linux & AIX are there main OSes for now, but how much work would it be to make Mac OS X on one of the faster boxes as a piece of the e-solutions push.
<strong>Though it appears that MacLuv fell on this data by accident (later saying that he intended that the UNIX server market is the relevant indicator, which it is not), it does seem that worldwide server market by units is the best indicator, so the numbers linked to Gartner's placing HP and Dell at the top, followed by IBM and then Sun, are an accurate indication pertinent to the iServe.</strong><hr></blockquote>
1. How does it appear that I fell on this data by accident? You've got evidence to prove this?
2. Why isn't the UNIX server market an "indicator"? It seems you haven't even bothered to check how the market is segmented.
3. What does this have to do with market penetration?
Tonton, I've bruised your ego. Sorry, it wasn't intentional. But build a bridge and get over it, please. At this rate you and Mr. Ed could run for office together.
<strong>I think it's obvious by now that MacLuv simply hates IBM, and is not willing to listen to any amount of reason, no matter how unanimous, to alter his opinion in the least.
He makes argumentative errors left and right, is guilty of numerous fallacies, especially personal attacks, special pleading and shifting the burden of proof, and he never holds himself accountable for his errors (he was the one who used the term "worldwide server market" and then blamed others for not pointing out his error earlier, as well as using the excuse "it was late"). Meanwhile, he argues over negligible semantics of posts made by others to distract from his own mistakes.
If there exists another as closed minded as he, such a person has yet to be seen.</strong><hr></blockquote>
If you want to point out specific examples that support this post, fine, but, like Ed M, you're doing nothing but harrasing me here. Politicians do this, are you running for office?
If you can point out where I've made errors then by all means do so. The fact is you can't, so you just attack me instead.
However, looking at IBM specifically, it would be ignorant not to notice that IBM (the market leader in terms of revenue due to their unarguable dominance in mid-to high-end sales) has determined to dive into this market head-first with the 970. To claim that this is insignificant for Apple, should they choose to incorporate the 970 into their own server solutions, would be foolish indeed.
</strong><hr></blockquote>
A conclusion reached by examining irrelevant information. You're suggesting Apple will have success with the 970 because of IBMs position in unrelated markets. You haven't tied the information together--in other words, Apple's success will not directly related to IBMs, unless the 970 shifts the industry away from x86. That's NEVER going to happen. x86 has already achieved critical mass. Anybody playing in that sandbox isn't worried about a new kid on the block.
[quote]<strong>
What I see is IBM, with a larger budget, world-renowned R&D, a more respected name, and superior technology, going up against Intel and AMD, who are clearly spending most of their efforts fighting over the consumer market on the tapped-out x86 architecture, and who have limited resources to commit to the mid-range server market.
</strong><hr></blockquote>
Apple is all about the "consumer" market. Even the Xserve seems positioned for the reach-out consumer ready to try a server solution.
The 970 is a commodoty just like any chip AMD or Intel makes. I cannot find any data that supports how much money they have devoted to making it an industry standard.
x86 isn't "tapped out". Technically it may be, but in the consumer market, that's irrelevant. This is about MONEY. The x86 ISA is here to stay for quite some time. As Windows is the industry standard OS, whatever direction Windows goes will be the direction supporting industry follows. The presence of the 970 will have no affect on this trend.
[quote]<strong>
I know where I would place my bets. IBM is going to kill in this market, and Apple is wise to hop on for the ride.
</strong><hr></blockquote>
The server market is a relationship market. It's not a consumer market. Regardless of which chip is in which machine, at the end of the day it depends upon who's shaken who's hand and who's kissed who's ass. That's how big business works.
[quote]<strong>
But this whole debate has migrated to concentrate on the server market. What about workstations?
</strong><hr></blockquote>
What about workstations?
This conversation migrated towards servers because I made one point that Apple sold 5,700 units and had a market penetration of 1.2%. Then some "consultant" boy posted information from 2001 about worldwide server domination without regards to market segmentation. So I pointed out that $$ within the UNIX market were dominated by Sun, and about five days ago HP became the leader. Then people start telling me that UNIX isn't UNIX, because of all the open source variations and whatnot.
Frankly, it's become a "whatever" issue at this point until a new case can be built around the evidence everyone has provided in this thread.
Comments
<strong>
If the X86 are currently better it's because of the great teams of engineers who develop them, not because the X86 code is inherently better : it's just the contrary. </strong><hr></blockquote>
Well, like I said, where there's a will, there's a way. What gave these engineers the incentive to push the x86 farther than what it was intended? Red Bull? Free GAP clothes? :cool:
<strong>
Well, like I said, where there's a will, there's a way. What gave these engineers the incentive to push the x86 farther than what it was intended? Red Bull? Free GAP clothes? :cool: </strong><hr></blockquote>
I just say that IBM has the will, Mot did not have it. The PPC 970 is the celeron version of the Power4. The Power5 is already on the way : there will be a future celeron or duron like version of this chip for the mac. This line of chip has a future.
Like all celerons like chips, the 970 chips does not represent a terrific amount of R&D at the exception of the two altivec units wich represant a great amount of R&D. But the job is already done, and less R&D will have to be done in the future developpements of the PPC 970.
In other way the power 4 has required a vast amount of R&D : it was a 4 years project developpement. The power line is a huge locomotive that drag behind him the PPC 970 line.
[ 12-08-2002: Message edited by: Powerdoc ]</p>
<strong>I don't even know what a low-end server market is.</strong><hr></blockquote>
But you still argue about it? Wow.
[quote]Originally posted by MacLuv:
<strong>Is that the entry level market that everyone with a clue talks about? Because that includes servers ~ $100,000.</strong><hr></blockquote>
Low end as in cheap, low cost servers that still are servers and not PC boxes running Linux. ~ $100,000.- servers are now low end since they pack quite some performance already. You could probably abstract it to the level that low end servers are 1-4 CPUs. But then again, for people deadling with servers worth $10,000,000 low end might be a 100,000 server.
At least from how I understood it.
<strong>In the contrary the power 4 represented a great amount of R&D.</strong><hr></blockquote>
Which, in turn, IBM does not want to be in vain. Hence this means they have longer-reaching plans for the platform.
<strong>
Which, in turn, IBM does not want to be in vain. Hence this means they have longer-reaching plans for the platform.</strong><hr></blockquote>
I think they have plans for the high performance servers based upon power chips and the low end ones based upon the PPC 970. You will notice is that if Apple was not concerned by the PPC 970 they would not have developped a SIMD unit Altivec compatible. A simple core power 4 without altivec unit would have been simplier to produce and design.
I think that producing PPC 970 is a way to recover a certain amount of R&D they spent on the power line.
<strong>
But you still argue about it? Wow.
</strong><hr></blockquote>
I was being facetious. Since you guys want me to specify "x86 ISA" rather than "x86 processor" perhaps you should use the term "entry-level" instead of "low-end". Low-end generally means cheap as in manufacture, not price. Actually I don't care which you use, as long as you know.
I've answered the questions to segmenting the server market with <a href="http://forums.appleinsider.com/cgi-bin/ultimatebb.cgi?ubb=get_topic&f=1&t=002736#000005"; target="_blank">this post.</a>
The immense R&D is part of the reason that Power4 is so expensive (other factors being low volume compared to desktop processors and the supporting arhcitecture being beefy).
<strong>amorph...
so you're saying that Apple is entering the PC server market?
</strong><hr></blockquote>
No, xype got it right: I'm saying that they're entering the low-end, UNIX-like server market. There are entries here from just about everyone, some running on platforms other than x86 (although x86 is common here), and there is lots of competition from the Windows server market, which has a large low-end presence.
It's basically the sub-$10K market. That's the neatest way to identify it. But if you want to segregate by OS also, then Apple is competing with Linux, BSD, and Solaris/x86 solutions under $10K.
[ 12-08-2002: Message edited by: Amorph ]</p>
[QB]
No, xype got it right: I'm saying that they're entering the low-end, UNIX-like server market. There are entries here from just about everyone, some running on platforms other than x86 (although x86 is common here), and there is lots of competition from the Windows server market, which has a large low-end presence.
It's basically the sub-$10K market. That's the neatest way to identify it. But if you want to segregate by OS also, then Apple is competing with Linux, BSD, and Solaris/x86 solutions under $10K.
[Oh, by the way: I suppose "entry level" would be more precise, but "low end" is still accurate. You pay for robust hardware, so the best you'll see sub-$10K are consumer-grade components like the x86, with some redundancy here and there.]
Whether Apple is entering 'low end', 'low price', 'UNIX' or 'UNIX-like' - I don't care. They're entering the server market. Like Steve Jobs said, they're entering it humbly.
Why are we discussing IBM's and Sun Microsystems' big iron servers when talking about what Apple should or shouldn't do? I don't see Apple in _that_ market in the near future (the next five years). And if I want a stable and fast web/file/mail/print/whatever-server for 'small' to 'medium', the Xserve is a nice alternative to what's around, if you consider price, usability, stability and power.
I'm sure the PowerPC 970 will be the better option for the Xserve than a Motorola chip, but an X86 chip (define that yourself) isn't a better option, unless you suggest that Apple would go X86 for the other product lines, too. But that's basically out of question for the near future (2003).
What Apple wants, in my opinion, is to build a strong option for businesses to go all Mac from server to desktop to notebook. They have a strong selling point there, but they're not well accepted in that market. This has nothing to do with Intel or AMD processors, but rather with how the company is viewed from a business angle. Apple looks like a consumer company to businesses. Very much like Sony. Xserve has made a good entry into the market (I don't really CARE about exact market share figures, this thread shows how irrelevant those statistics are, because it very much depends on how you look at them.). It's gotten good reviews, it fares well in its price/power segment and people start to look at Apple differently. Maybe in a year or two, Apple will be taking a bite out of Dell for business sales. And business people could care less about what processor is in a machine. ROI is important. TCO is important. You don't see many businesses upgrading processors every other months, only because Intel claims that 'the internet is faster with MMX' or anything like that. But they _do_ see that Microsoft is a one-way street. They _do_ see that Microsoft is starting to behave strangely with licenses. And Microsoft pissing off people is an opportunity for Apple.
"I'm not sure what their deal is, but we should be seeing 1.2GHz to 1.6GHz from the die shrink."<hr></blockquote>
The 64 million dollar question is WHEN. There have been very few rumors regarding this lately.
OFF TOPIC
[quote]Originally posted by MacLuv:
<strong>
The whole point of getting into servers and marketshare, etc, was to talk about Apple's market penetration into the server world.
</strong><hr></blockquote>
My bad. I mistakenly thought your original comments regarding IBM's small market share and slight decline in UNIX servers was in response to positive arguments made for Apple using an IBM cpu, because of IBM's large size and ability to spend billions on research.
After your statement of facts, it seemed the argument changed from IBM's commitment/ability to design/develop/market a competitive cpu to Apple's initial server offerings. Regrettably, this confused me.
many many spelling errors
[ 12-08-2002: Message edited by: rickag ]</p>
<strong>I was being facetious. Since you guys want me to specify "x86 ISA" rather than "x86 processor" </strong><hr></blockquote>
That was just me being pedantic, actually. I don't speak for anybody else here, and nobody else here speaks for me.
All this talk of UNIX vs Linux markets is really a little bizarre, IMO. For starters, Unix isn't one operating system -- there is about one flavour per vendor. They are all more or less compatible, except that they run on different processors. Then there is the FreeBSD crowd, the Linux crowd, and now Mac OS X. All of these are quite similar, but they are all different in various details. A lot of software can just be recompiled across the various flavours, and administration/user knowledge is largely transferable but its not like Windows or Macintosh in that you can buy shrink wrapped software and expect it to run on any Unix-like OS. In the server market that is acceptable (and perhaps even expected), but its hopeless in the consumer desktop market. In the server market, therefore, IMO the operating systems should be lumped into Windows and Unix-like. In the desktop market its really just Windows and Macintosh (with the various versioning issues inherent in each of those). Anybody else is in the geek market and is so individualistic that it doesn't matter.
I'm amazed that there is anybody left who doesn't believe we'll see a 970-based Macintosh from Apple at some point next year. That processor officially exists, it has VMX, it is aimed at the desktop, Steve Jobs himself has said that they like what IBM has lined up, etc.
There might be another G4 revision (0.13 micron is strongly hinted at), but who knows how fast it'll be. Moto's scaling problems could persist and 1.3 GHz might be the limit despite their earlier expectation that they could claw their way up to 1.8 GHz eventually. Sure the 0.18 -> 0.13 process shrink is expected to deliver a 30% performance boost, but that is a guideline not a sure thing. Regardless of what they actually get out of it, the G4 is likely to continue to drive the consumer & notebook lines for a while yet... at least until the 0.09-micron 970 arrives.
IBM has built a 64-bit PowerPC before, remember -- the 620 and 630 were designed and produced around 1997 and used by IBM's AS/400 group IIRC. The POWER3 & POWER4 were both 64-bit as well. Lots of experience and tools to be had there. PowerPC was designed from day 1 to have a 64-bit version that was backward compatible with the 32-bit version. x86 wasn't even designed to have a 16-bit version!!! Its been one hack on top of another one.
The x86 market will find its way despite themselves... they've got enough money to spend their way to a solution. It doesn't take willpower to build an advanced processor, it just takes money. Lots of it. Intel has it, IBM has it (and a better starting position), Motorola doesn't have it, AMD might have it. Who are you going to bet on?
<strong>Like Steve Jobs said, they're entering it humbly.
Why are we discussing IBM's and Sun Microsystems' big iron servers when talking about what Apple should or shouldn't do? I don't see Apple in _that_ market in the near future (the next five years).</strong><hr></blockquote>
To be honest Apple doesn't have the resources to enter any bigger a server Market than what can be achieved with up to 4-way CPU servers (lowe end) because the higher segments demand, like Amorph said, quality that has to be proven and that costs a lot. IBM has a reputation of a few decades they can build upon and that's ecactly the reason people with loads of cash to spend but IBM hardware - because for 10.000.000 you better get something that works - and works the way you want. Which is also the reason SGI stayed afloat, because they went back to do what they know best - powerful computing. And not "imagine a Beowulf cluster of these" wannabe power-computing.
The Rack-mountable approach Apple is taking is perfectly fine because it's a market where they could in theory live off their current customer base and compete with cheap Linux boxes. Plus the R&D cost is not really high so the risk is far lower than going head to head with IBM/SGI/Sun for the real server market. Just like their RAID system is neat for the Apple market, but no way they could ever challenge EMC and Hitachi.
Apple is playing some well tought out moves lately and if it weren't for their bus/cpu cripple they'd be in a really good position.
<strong>The x86 market will find its way despite themselves... they've got enough money to spend their way to a solution. It doesn't take willpower to build an advanced processor, it just takes money.</strong><hr></blockquote>
I will argue that - Intel really doesn't have enough willpower to go for something better, they were really content with simply pushing mhz for a long time. Besides, the goal behind most of x86 hardware designs is low cost cheapcheap and that alone will make the platform stay unappealing for most "serious" computer people.
PowerPC always seemed to me like a platform where the functionality was given higher priority than cost/mhz, which is a good thing. It may have not always been in Apple's best interest (IBM servers, Motorola embedded) but if IBM wants to go the workstation/desktop route, they have a really solid base to build upon and a better starting point than the x86 world. Heck, I doubt Apple will ever have problems because or fubared IRQs or because the motherboard will have conflicts with their graphics card. Unless they switch to AMD.
<strong>
That was just me being pedantic, actually. I don't speak for anybody else here, and nobody else here speaks for me.
</strong><hr></blockquote>
No, it's fine that you pointed out that perhaps I should be referring to "x86 ISA" because--as I went back to examine if it would have any relevance to my agruments--I learned more about how the instruction set would impact the future. For example, as the x86 ISA "moves forward", there's always going to be a billion dollar industry for x86 legacy support. To an idealist this is absurd, to an entrepreneur this is opportunity.
Regarding "Linux vs Unix" variations, etc... industry analysts have to draw the line somewhere. They can't sit down and nitpick over which side Kirk's hair was parted in episode #37. This doesn't mean I don't agree with what you and others are saying--that there are "grey" areas of distinction--but for business purposes the market has already been segmented. I didn't do it, Gartner's & others have. There are plenty of subs that may be spitting hairs with markets right now but I haven't bothered to check. In that the Xserve is under $10K and everyone is claiming that this is a defined market, I ask for help finding out who has defined that market... I don't have the time right now.
[ 12-08-2002: Message edited by: MacLuv ]</p>
<strong>It's basically the sub-$10K market. </strong><hr></blockquote>
I could see them reaching up to ~ $15,000. But only if the XServe & RAID version work well. (Initial reviews & reports on the XServe appear to be favorable) Some of the cheap blade server approaches would work well with a low heat chip, Apple could do one well, and the bioinformics groups (at the least) would love them. It seems like that the few big problems that are easily vectorizible have a boundless appreciation for computing power. More AV units in a box = good thing.
The part I'm curious about is whether the relationship between IBM and Apple might end up being close enough for IBM to offer something that Mac OS X can run on easily. Linux & AIX are there main OSes for now, but how much work would it be to make Mac OS X on one of the faster boxes as a piece of the e-solutions push.
<strong>Though it appears that MacLuv fell on this data by accident (later saying that he intended that the UNIX server market is the relevant indicator, which it is not), it does seem that worldwide server market by units is the best indicator, so the numbers linked to Gartner's placing HP and Dell at the top, followed by IBM and then Sun, are an accurate indication pertinent to the iServe.</strong><hr></blockquote>
1. How does it appear that I fell on this data by accident? You've got evidence to prove this?
2. Why isn't the UNIX server market an "indicator"? It seems you haven't even bothered to check how the market is segmented.
3. What does this have to do with market penetration?
Tonton, I've bruised your ego. Sorry, it wasn't intentional. But build a bridge and get over it, please. At this rate you and Mr. Ed could run for office together.
<img src="graemlins/oyvey.gif" border="0" alt="[No]" />
[ 12-09-2002: Message edited by: MacLuv ]</p>
<strong>I think it's obvious by now that MacLuv simply hates IBM, and is not willing to listen to any amount of reason, no matter how unanimous, to alter his opinion in the least.
He makes argumentative errors left and right, is guilty of numerous fallacies, especially personal attacks, special pleading and shifting the burden of proof, and he never holds himself accountable for his errors (he was the one who used the term "worldwide server market" and then blamed others for not pointing out his error earlier, as well as using the excuse "it was late"). Meanwhile, he argues over negligible semantics of posts made by others to distract from his own mistakes.
If there exists another as closed minded as he, such a person has yet to be seen.</strong><hr></blockquote>
If you want to point out specific examples that support this post, fine, but, like Ed M, you're doing nothing but harrasing me here. Politicians do this, are you running for office?
If you can point out where I've made errors then by all means do so. The fact is you can't, so you just attack me instead.
Get over it, bud.
<img src="graemlins/oyvey.gif" border="0" alt="[No]" />
<strong>
However, looking at IBM specifically, it would be ignorant not to notice that IBM (the market leader in terms of revenue due to their unarguable dominance in mid-to high-end sales) has determined to dive into this market head-first with the 970. To claim that this is insignificant for Apple, should they choose to incorporate the 970 into their own server solutions, would be foolish indeed.
</strong><hr></blockquote>
A conclusion reached by examining irrelevant information. You're suggesting Apple will have success with the 970 because of IBMs position in unrelated markets. You haven't tied the information together--in other words, Apple's success will not directly related to IBMs, unless the 970 shifts the industry away from x86. That's NEVER going to happen. x86 has already achieved critical mass. Anybody playing in that sandbox isn't worried about a new kid on the block.
[quote]<strong>
What I see is IBM, with a larger budget, world-renowned R&D, a more respected name, and superior technology, going up against Intel and AMD, who are clearly spending most of their efforts fighting over the consumer market on the tapped-out x86 architecture, and who have limited resources to commit to the mid-range server market.
</strong><hr></blockquote>
Apple is all about the "consumer" market. Even the Xserve seems positioned for the reach-out consumer ready to try a server solution.
The 970 is a commodoty just like any chip AMD or Intel makes. I cannot find any data that supports how much money they have devoted to making it an industry standard.
x86 isn't "tapped out". Technically it may be, but in the consumer market, that's irrelevant. This is about MONEY. The x86 ISA is here to stay for quite some time. As Windows is the industry standard OS, whatever direction Windows goes will be the direction supporting industry follows. The presence of the 970 will have no affect on this trend.
[quote]<strong>
I know where I would place my bets. IBM is going to kill in this market, and Apple is wise to hop on for the ride.
</strong><hr></blockquote>
The server market is a relationship market. It's not a consumer market. Regardless of which chip is in which machine, at the end of the day it depends upon who's shaken who's hand and who's kissed who's ass. That's how big business works.
[quote]<strong>
But this whole debate has migrated to concentrate on the server market. What about workstations?
</strong><hr></blockquote>
What about workstations?
This conversation migrated towards servers because I made one point that Apple sold 5,700 units and had a market penetration of 1.2%. Then some "consultant" boy posted information from 2001 about worldwide server domination without regards to market segmentation. So I pointed out that $$ within the UNIX market were dominated by Sun, and about five days ago HP became the leader. Then people start telling me that UNIX isn't UNIX, because of all the open source variations and whatnot.
Frankly, it's become a "whatever" issue at this point until a new case can be built around the evidence everyone has provided in this thread.
<img src="graemlins/smokin.gif" border="0" alt="[Chilling]" />
[ 12-09-2002: Message edited by: MacLuv ]</p>