Okay. If I were IBM, having pumped hordes of cash into developing a desktop Power4 chip with a new plant to manufacture them, i would be selling the chips to as many customers as possible. Furthermore, from an overal business strategy, it makes sense for IBM to expand the influence of PPC - they own it, and they don't have to buy chips from anyone else in order to use it. If Apple uses the same PPC chips as they do -power to them! Different markets. Linux is not a home computer for the masses - the Mac is.
nuff said.
</strong><hr></blockquote>
Well put, Mactivist. I think I was somewhat unclear in my post, which let to your critique. Before I start my reply, let me tell you though, that I appreciated your response very much. Your points were all valid.
I did not mean to say that IBM won't offer its new processors to Apple. IBM will happily supply processors to whoever has the money to buy them.
What I meant to say is, that IBM did certainly not develop this processor with Apple as their premier customer in mind. Hence I suggest that we try to discover IBM's main impetus for devoloping this new chip.
Making it as a Linux platform, with Apple and others as secondary customers for more volume/profit, made the most sense to me. The better we understand IBM, the better we can understand IBM's impact on Apple.
What I meant to say is, that IBM did certainly not develop this processor with Apple as their premier customer in mind.
</strong><hr></blockquote>
If IBM didn't develop it specifically for Apple, I am sure Apple was involved, so the chip meets Apple's needs. They wouldn't just hope for that; they would work with Apple. Also, when IBM starts making these in high volume, they want to start shipping right away. The only way to do that is if Apple has a product designed and ready for production, and uses them. Yes, Apple has been testing this G5 for some time now.
This processor didn't catch Apple by surprise. If it did, it means that for the last three years, Steve Jobs has just been hoping Motorola would get things right someday. I believe Steve knew the likelyhood of that was small, and he was looking for another source for the G5.
So IBM is laying off 15k people will that affect powerpc development?</strong><hr></blockquote>
It's mostly (or all) in their services division. Plus, if this chip was designed for Apple, then Apple has already placed the order. The only thing left is for IBM to fulfill that order.
<strong>Yes indeed, time is free. And it is the industry standard.</strong><hr></blockquote>
If you say time is free then you obviously don't have a job yet. Nor anything serious to do with it. And if by industry standard you mean Linux then I say "Ha-ha!" - or are you talking about some othe Linux than the one with 39 different glibc versions, whose software comes with various formats (was that rpm or something), which you need half a day to set up as a proxy server, which has at least two different desktop environments (each with its own software) and where you can't use any exotic piece of hardware? Is that what standards are like?
What I meant to say is, that IBM did certainly not develop this processor with Apple as their premier customer in mind. Hence I suggest that we try to discover IBM's main impetus for devoloping this new chip.
</strong><hr></blockquote>
Unless you have inside information, there is no way you can know that.
The only thing we can reasonably assume is that IBM already has a customer for this processor. The customer may be Apple. The customer may be IBM. The customer may even be Amiga. However, some of these are more likely than others.
Without knowing the exact specs of the chips, we can say the following:
Apple: This chip is pretty much EXACTLY what Apple needs. IBM could not have done a better job at giving Apple backward compatability while at the same time providing a nice looking future. Couple this with Steve-O buying every high-end graphics package he can find. Add to that Steve-O's knowledge that the current Mac's are NOT up to the task when it comes to running those packages. Throw in that Motorola doesn't appear to be offering anything new. Tie it up with Steve-O not being so stupid (probably not) as to buy all that kick-arse software without knowing where he's gonna get some kick-arse hardware to run it on. Steve-O already KNOWS what the kick-arse hardware is going to look like. He already KNOWS what processor will be in the thing. He knew it quite a while ago, when he started his shopping spree. It's being designed (or finished?) as we speak.
IBM: What would IBM use this for? Every argument I hear is nothing but wild speculation. IBM MAY be getting into the PPC desktop market. With what OS? Well, maybe Linux. Certainly not OS/2. WinXP? Ha! People can get a really decent Intel/AMD PC and Linux for FREE today. People can get this TODAY. And they're NOT. NOBODY IS BUYING LINUX for the desktop. So we're to believe that IBM is going to invest BILLIONS in designing a new PPC chip, create a new PPC desktop line, and try to sell it as a Linux package? How would their product differ from an Intel/AMD solution? Why would it be any better? Why would people who don't want Linux on their desktop today want it from IBM tomorrow?
Perhaps IBM is simply developing the chip first and hoping someone will buy it? Who? Apple? Good potential choice. IBM? They'd already know, wouldn't they? Amiga? They wouldn't purchase enough to pay the electric bill. Sun? SGI? Sony? Nintindo?
SOMEONE has already placed an order for this chip. The most logical of all choices is Apple. This chip will be in a PowerMac within 12 months. No question.
1) IBM's needs PowerPC workstations so that people develop software for their profitable midrange servers.
2) Very few non-AIX devs buy these workstations. The market is so tiny that it could almost afford the existing POWER4.
3) Any new IBM product will run Linux as a matter of policy.
4) But IBM isn't stupid and realizes that they have very little chance positioning these workstaions against "LinTel", not to mention IBM's Intel workstations are already top-notch.
4a) RedHat doesn't sell CPUs.
5) PREP/CHRP is dead and decayed and forgotten. PowerPC will not be positioned against Intel/AMD in the commodity market, there is 0 industry interest.
5) Therefore, it wouldn't make sense for IBM to develop this chip unless they had a major customer besides themselves.
IBM has several lines of machines that desperately need a new processor -- their desktop workstations, low-end servers, AS/400 business servers, and very large scale SMP projects (which need a lower power chip than the POWER4). Many of these are currently using ancient 604e processors and are hugely under powered. This alone is a good reason for IBM to develop such a processor. This is the same reason that Sun continues to build new chips, I don't see what is so hard to understand about this (much less argue about). If Apple can use it to then IBM can recoup the investment costs much faster thanks to the 1-2 million chips per year Apple will need (esp. if they ship duals still).
Adding VMX to the design is pretty much a no-brainer these days -- every other major chip design has a SIMD unit, and AltiVec is the best-of-breed and IBM holds the rights to build it. Not only does it benefit their machines (and they have been talking about this kind of enhancement for a year or two now), but it makes Apple a strong prospective customer. At $250/chip that is at least a quarter billion per year.
And I wish people would keep going on and on about how expensive this chip will be because it is derived from a POWER4. Don't you get it?! Its not a POWER4!! They will take the 30-million transistor core, surround it with infrastructure appropriate to a 0.1 or 0.13 micron desktop processor, and interface it to the machine with a next generation desktop bus interface (RIO, HT, or something like that). It could be less expensive than the Pentium4 and Althon, not to mention next year's monster x86 chips. This design will be carefully targeted to deliver an appropriate mix of performance, price, yield, and power consumption... with the emphasis (finally) on performance. The Pentium4 is somewhere around 60 million transistors, and next year's standards are going to be even higher. Current graphic chips are over 100 million. This new PowerPC will simply be the first in a few years that is actually designed to compete with the Intel desktop chips... the last was the 604e, which (as mentioned above) IBM is still using.
The only two real questions remaining are: what will the technical details be, and when can Apple start shipping machines based on it? I'm with Moki on the second point -- latter half of next year. As to the first, I guess we'll have to wait until the next leak or October's MDF.
<strong>The question is will this "new" chip go into PowerMac pro-sumer models or is there a new $8k workstation coming out from Apple</strong><hr></blockquote>
And the answer is obvious: It will replace the G4 in the PowerMac and Xserve. It wouldn't be a "desktop and entry-level server" CPU if Apple had to charge $8K per box to maintain a +/-30% profit margin, now would it?
IBM isn't stupid, they're not going to develop a desktop CPU and sell it at POWER4 prices. That would defeat the whole purpose of developing a new line of 64-bit PowerPCs, only to fill the same market niche as the POWER4 currently does.
<strong>Its not a POWER4!! They will take the 30-million transistor core, surround it with infrastructure appropriate to a 0.1 or 0.13 micron desktop processor, and interface it to the machine with a next generation desktop bus interface (RIO, HT, or something like that).</strong><hr></blockquote>
My question for those of you more knowledgable about the technical stuff is this: How much of the POWER4's legendary performance is due to the 30-million transistor core and how much of it is due to the infrastructure that surrounds that core.
So IBM is laying off 15k people will that affect powerpc development?</strong><hr></blockquote>
As reported by management within IBM, "The only publication to get the news release correct, was the Wall Street Journal". IBM has laid off 15,000 people (this last June), from all areas within IBM. It was approximately 5% of the total number of IBM employees. I heard it directly from IBM upper management just yesterday. The news services that say IBM will lay off another 15K people are incorrect in their interpretation of IBM's news release.
Apple's G5 may come from IBM but it may be totally irrelevant to IBM's plans.
Given that IBM really does well in the server market and they are pressing ahead with Linux for everything they sell, it makes sense that they are targeting this chip family towards mainstream low-end serving duties. In the server space Linux does really well with a signicant and growing share of the total market which is far larger than the small declining desktop market Apple has.
OSX performance is really sluggish right now especially for serving duties compared to Linux. In fact, its a huge slug for file serving, DNS, Samba, FTP, SQL, and Apache. There may be other but those are the services where we've compared the two.
But then again Apple is not going after the server market with OSX/G4 with any real effort. Which is good since every Mac user with a few exceptions is happy with what they perceive to be is a "Better User Experience" on slow hardware.
There are indications that Apple have been having close discussions with AMD and nVidia, I suspect the nVidia discussions, if anything other than talks about graphic cards, is about integrating an nVidia graphics unit into Apple's system controllers.
What do AMD have that Apple would like? Not their fab, it's fully used anyway, nor their processors (too foreign to PPC), but they have a very nice interface portion of their Hammer chips. The Hammer chips have a set of Hypertransport links, onboard memory controllers, and a switch to connect them, this section could relatively easily be put around a different core (eg: a Power4 core + SIMD, or maybe 2) to give a very nice chip. This would also considerably speed up the development of the chip, any thoughts.
<strong>Apple's G5 may come from IBM but it may be totally irrelevant to IBM's plans.
Given that IBM really does well in the server market and they are pressing ahead with Linux for everything they sell, it makes sense that they are targeting this chip family towards mainstream low-end serving duties. In the server space Linux does really well with a signicant and growing share of the total market which is far larger than the small declining desktop market Apple has. </strong><hr></blockquote>
Actually, this isn't true. Linux's penetration into the server space is actually on the decline -- at least temporarily -- and its desktop installations aren't even on the radar -- Apple came in with Mac OS X, and almost overnight surprassed Linux in terms of installed seats.
Linux for geeks and sysadmins to run on their servers is one thing; running it on the desktop is something it has failed miserably at, and rightly so: it sucks as a desktop OS. Do you honestly believe that there is a larger market for servers than there is for desktop computers? What would they be serving to, if that were the case?
[quote] Do you honestly believe that there is a larger market for servers than there is for desktop computers? What would they be serving to, if that were the case?<hr></blockquote>
What, you never heard of "self-serve?"
I think it's clear that these processors are aimed at Apple's desktop market. IBM already deploys Linux on commodity x86 platforms; what's the point of developing an entire new architecture to run the same old stuff?
If it ain't broke, don't fix it, so the saying goes.
Now Apple's pro processor situation- that's broke. I think IBM's fixin' to change it.
There comes a point when slow hardware does affect the 'superior' user experience.
eg.
I thought I'd try the dual giggers again. Running Photoshop. Maybe it was just me last time. Maybe they were really fast.
Tried throwing around 8 by 11 inch 300 dpi images in PS7 on 'X'.
512 megs of ram. No small amount.
My opinion?
Pauseworthy. PS7 on 'X' drags its feet. The dual gigger crawls through many of the Photoshop filters. Not enough that you'd go away and make a cup of tea...but yeesh. This is Apple's flagship? (Well, it was 8 months ago...and still is...seeing as the tiny bump cheese grater won't be shipping until end of September.)
Not impressed.
My point? My wife's 384 megs/600 G3 iBook running Photoshop 4 on Classic doesn't suffer in a perceptional comparison. You may snigger at that...but...I'm not.
Apple are innovating with OS, Software and design.
Everything but the actual specs.
I don't think we're going to see the kind of performance to run X or its applications properly until we leave the G4 behind. It just can't hack it. It was designed to compete with low mhz Pentium 3...which are now part of Intel's budget range.
Somebody calls the dual giggers 'Slow Silvers' on these boards. He's right, y'know.
Having said that. I've recently got my wife's iBook onto a DSL connection and I haven't been able to stop using it for the internet! 'X' is great. 'Anti Aliased' type...gel buttons on Yahoo pages...its...beautiful. I've got it mirrored through a 21inch monitor (which used to 'mirror' my PowerMac 200mhz clone...) and 'X' looks beautiful. It's a work of art.
My real point?
I'll concede that you don't have to have a 'G5' to enjoy 'X' now...but to do everything Apple is trying to do now...JUSTICE...then we're going to need that G5.
It's the last piece of the puzzle.
I believe now that the final piece will come from IBM.
Lemon Bon Bon
Enjoying his wife's iBook. Still 'power'Mac less though. Come on Apple, release the darn G5 already...
[[[Apple's G5 may come from IBM but it may be totally irrelevant to IBM's plans.
Given that IBM really does well in the server market and they are pressing ahead with Linux for everything they sell, it makes sense that they are targeting this chip family towards mainstream low-end serving duties. In the server space Linux does really well with a significant and growing share of the total market which is far larger than the small declining desktop market Apple has. ]]]
This is totally incorrect. Forget market share. Look at the numbers of actual "users". Apple's user base is GROWING. So stop trolling.
IBM is NOT going to try to bring LINUX into the consumer mass-market. OS X is more developed, usable and more robust and ALREADY in the mass-market for consumers.
As a matter of FACT, Apple is now the LARGEST distributor of a commercial-grad UNIX-based operating system.
David K. Every said it best...
[[[OSX is already the unit leader in terms of a UNIX distribution, BUT... Other people that rely on UNIX (corporate entities et. al.) are learning that IT IS ALREADY THE INTERFACE LEADER as well. And this is 12 months out of the gate. Jaguar looks to fix/improve many things... And in another 12-18 months, I think it will be the standard by which most UNIXXES are measured... OSX delivers today on what LINUX has been promising for 10 years... ]]] - DKE
IBM is obviously aware of this as well. Now how does that figure into the equation and Why this isn't a signal that IBM is going to try another attempt to win-over the desktop? Well, all you have to do is think about it...
What reason would IBM all of a sudden start to utilize AltiVec (or some other version of the original VMX design)? The embedded market? Unlikely, IBM stated that they developed a CPU *for the desltop*. Now, why would they do that? What other OS runs on PPC, LINUX? OK, I'll buy that.. But what would LINUX need AltiVec for? OK, maybe there are a *few* obscure, not-often-used-by-average-people apps. Still, there are NONE that are used in the mainstream.
Would these machines be running a PPC-LINUX version of Photoshop with AltiVec optimizations? No.
What about a LINUX version of Lightwave or Maya with AltiVec optimizations? Nope.
Office and AutoCAD for LINUX? No.
Will there suddenly be a massive rush of developers to code any of this stuff for IBM? I doubt it.
That leads to only one, logical conclusion...
Apple *already* has the OS that utilizes both the PPC and AltiVec quite well. There are PLENTY of apps already in place that have been utilizing these technologies for a while now and the brand is well-known. For developers, It is cheaper to make it work on OSX than other UNIXXEN, and OS X already has a larger installed base.
And as for people who think that the vector unit on the IBM cpu *isn't* AltiVec, let me remind you that AltiVec is just as much a technology of IBM as it is to Apple and Motorola. IBM wouldn't need special licensing. Furthermore, It's unlikely to believe that IBM would go through all the trouble to develop a completely new vector unit when they have a perfectly good one already available. Additionally, why would they bother to develop new tools to work with the new vector unit and how much training would be involved trying to get developers accustomed to using the technology properly?
Apple has the OS in place AND the commercial apps that LINUX simply doesn't have. OS X has a UI that's much further along and continuing to make strides.
Comments
<strong>
Okay. If I were IBM, having pumped hordes of cash into developing a desktop Power4 chip with a new plant to manufacture them, i would be selling the chips to as many customers as possible. Furthermore, from an overal business strategy, it makes sense for IBM to expand the influence of PPC - they own it, and they don't have to buy chips from anyone else in order to use it. If Apple uses the same PPC chips as they do -power to them! Different markets. Linux is not a home computer for the masses - the Mac is.
nuff said.
Well put, Mactivist. I think I was somewhat unclear in my post, which let to your critique. Before I start my reply, let me tell you though, that I appreciated your response very much. Your points were all valid.
I did not mean to say that IBM won't offer its new processors to Apple. IBM will happily supply processors to whoever has the money to buy them.
What I meant to say is, that IBM did certainly not develop this processor with Apple as their premier customer in mind. Hence I suggest that we try to discover IBM's main impetus for devoloping this new chip.
Making it as a Linux platform, with Apple and others as secondary customers for more volume/profit, made the most sense to me. The better we understand IBM, the better we can understand IBM's impact on Apple.
<strong>
What I meant to say is, that IBM did certainly not develop this processor with Apple as their premier customer in mind.
</strong><hr></blockquote>
If IBM didn't develop it specifically for Apple, I am sure Apple was involved, so the chip meets Apple's needs. They wouldn't just hope for that; they would work with Apple. Also, when IBM starts making these in high volume, they want to start shipping right away. The only way to do that is if Apple has a product designed and ready for production, and uses them. Yes, Apple has been testing this G5 for some time now.
This processor didn't catch Apple by surprise. If it did, it means that for the last three years, Steve Jobs has just been hoping Motorola would get things right someday. I believe Steve knew the likelyhood of that was small, and he was looking for another source for the G5.
[ 08-17-2002: Message edited by: snoopy ]</p>
<strong>this thread has gone off topic, time to lock it and restart.
G-News</strong><hr></blockquote>
let the admins do their job.
So IBM is laying off 15k people will that affect powerpc development?
<strong>
So IBM is laying off 15k people will that affect powerpc development?</strong><hr></blockquote>
It's mostly (or all) in their services division. Plus, if this chip was designed for Apple, then Apple has already placed the order. The only thing left is for IBM to fulfill that order.
<strong>Yes indeed, time is free. And it is the industry standard.</strong><hr></blockquote>
If you say time is free then you obviously don't have a job yet. Nor anything serious to do with it. And if by industry standard you mean Linux then I say "Ha-ha!" - or are you talking about some othe Linux than the one with 39 different glibc versions, whose software comes with various formats (was that rpm or something), which you need half a day to set up as a proxy server, which has at least two different desktop environments (each with its own software) and where you can't use any exotic piece of hardware? Is that what standards are like?
<strong>
What I meant to say is, that IBM did certainly not develop this processor with Apple as their premier customer in mind. Hence I suggest that we try to discover IBM's main impetus for devoloping this new chip.
</strong><hr></blockquote>
Unless you have inside information, there is no way you can know that.
The only thing we can reasonably assume is that IBM already has a customer for this processor. The customer may be Apple. The customer may be IBM. The customer may even be Amiga. However, some of these are more likely than others.
Without knowing the exact specs of the chips, we can say the following:
Apple: This chip is pretty much EXACTLY what Apple needs. IBM could not have done a better job at giving Apple backward compatability while at the same time providing a nice looking future. Couple this with Steve-O buying every high-end graphics package he can find. Add to that Steve-O's knowledge that the current Mac's are NOT up to the task when it comes to running those packages. Throw in that Motorola doesn't appear to be offering anything new. Tie it up with Steve-O not being so stupid (probably not) as to buy all that kick-arse software without knowing where he's gonna get some kick-arse hardware to run it on. Steve-O already KNOWS what the kick-arse hardware is going to look like. He already KNOWS what processor will be in the thing. He knew it quite a while ago, when he started his shopping spree. It's being designed (or finished?) as we speak.
IBM: What would IBM use this for? Every argument I hear is nothing but wild speculation. IBM MAY be getting into the PPC desktop market. With what OS? Well, maybe Linux. Certainly not OS/2. WinXP? Ha! People can get a really decent Intel/AMD PC and Linux for FREE today. People can get this TODAY. And they're NOT. NOBODY IS BUYING LINUX for the desktop. So we're to believe that IBM is going to invest BILLIONS in designing a new PPC chip, create a new PPC desktop line, and try to sell it as a Linux package? How would their product differ from an Intel/AMD solution? Why would it be any better? Why would people who don't want Linux on their desktop today want it from IBM tomorrow?
Amiga: <img src="graemlins/oyvey.gif" border="0" alt="[No]" />
Any others?
Perhaps IBM is simply developing the chip first and hoping someone will buy it? Who? Apple? Good potential choice. IBM? They'd already know, wouldn't they? Amiga? They wouldn't purchase enough to pay the electric bill. Sun? SGI? Sony? Nintindo?
SOMEONE has already placed an order for this chip. The most logical of all choices is Apple. This chip will be in a PowerMac within 12 months. No question.
Look at it from IBM's point of view.
1) IBM's needs PowerPC workstations so that people develop software for their profitable midrange servers.
2) Very few non-AIX devs buy these workstations. The market is so tiny that it could almost afford the existing POWER4.
3) Any new IBM product will run Linux as a matter of policy.
4) But IBM isn't stupid and realizes that they have very little chance positioning these workstaions against "LinTel", not to mention IBM's Intel workstations are already top-notch.
4a) RedHat doesn't sell CPUs.
5) PREP/CHRP is dead and decayed and forgotten. PowerPC will not be positioned against Intel/AMD in the commodity market, there is 0 industry interest.
5) Therefore, it wouldn't make sense for IBM to develop this chip unless they had a major customer besides themselves.
Adding VMX to the design is pretty much a no-brainer these days -- every other major chip design has a SIMD unit, and AltiVec is the best-of-breed and IBM holds the rights to build it. Not only does it benefit their machines (and they have been talking about this kind of enhancement for a year or two now), but it makes Apple a strong prospective customer. At $250/chip that is at least a quarter billion per year.
And I wish people would keep going on and on about how expensive this chip will be because it is derived from a POWER4. Don't you get it?! Its not a POWER4!! They will take the 30-million transistor core, surround it with infrastructure appropriate to a 0.1 or 0.13 micron desktop processor, and interface it to the machine with a next generation desktop bus interface (RIO, HT, or something like that). It could be less expensive than the Pentium4 and Althon, not to mention next year's monster x86 chips. This design will be carefully targeted to deliver an appropriate mix of performance, price, yield, and power consumption... with the emphasis (finally) on performance. The Pentium4 is somewhere around 60 million transistors, and next year's standards are going to be even higher. Current graphic chips are over 100 million. This new PowerPC will simply be the first in a few years that is actually designed to compete with the Intel desktop chips... the last was the 604e, which (as mentioned above) IBM is still using.
The only two real questions remaining are: what will the technical details be, and when can Apple start shipping machines based on it? I'm with Moki on the second point -- latter half of next year. As to the first, I guess we'll have to wait until the next leak or October's MDF.
<strong>The question is will this "new" chip go into PowerMac pro-sumer models or is there a new $8k workstation coming out from Apple</strong><hr></blockquote>
And the answer is obvious: It will replace the G4 in the PowerMac and Xserve. It wouldn't be a "desktop and entry-level server" CPU if Apple had to charge $8K per box to maintain a +/-30% profit margin, now would it?
IBM isn't stupid, they're not going to develop a desktop CPU and sell it at POWER4 prices. That would defeat the whole purpose of developing a new line of 64-bit PowerPCs, only to fill the same market niche as the POWER4 currently does.
<strong>Its not a POWER4!! They will take the 30-million transistor core, surround it with infrastructure appropriate to a 0.1 or 0.13 micron desktop processor, and interface it to the machine with a next generation desktop bus interface (RIO, HT, or something like that).</strong><hr></blockquote>
My question for those of you more knowledgable about the technical stuff is this: How much of the POWER4's legendary performance is due to the 30-million transistor core and how much of it is due to the infrastructure that surrounds that core.
<strong>
let the admins do their job.
So IBM is laying off 15k people will that affect powerpc development?</strong><hr></blockquote>
As reported by management within IBM, "The only publication to get the news release correct, was the Wall Street Journal". IBM has laid off 15,000 people (this last June), from all areas within IBM. It was approximately 5% of the total number of IBM employees. I heard it directly from IBM upper management just yesterday. The news services that say IBM will lay off another 15K people are incorrect in their interpretation of IBM's news release.
Given that IBM really does well in the server market and they are pressing ahead with Linux for everything they sell, it makes sense that they are targeting this chip family towards mainstream low-end serving duties. In the server space Linux does really well with a signicant and growing share of the total market which is far larger than the small declining desktop market Apple has.
OSX performance is really sluggish right now especially for serving duties compared to Linux. In fact, its a huge slug for file serving, DNS, Samba, FTP, SQL, and Apache. There may be other but those are the services where we've compared the two.
But then again Apple is not going after the server market with OSX/G4 with any real effort. Which is good since every Mac user with a few exceptions is happy with what they perceive to be is a "Better User Experience" on slow hardware.
There are indications that Apple have been having close discussions with AMD and nVidia, I suspect the nVidia discussions, if anything other than talks about graphic cards, is about integrating an nVidia graphics unit into Apple's system controllers.
What do AMD have that Apple would like? Not their fab, it's fully used anyway, nor their processors (too foreign to PPC), but they have a very nice interface portion of their Hammer chips. The Hammer chips have a set of Hypertransport links, onboard memory controllers, and a switch to connect them, this section could relatively easily be put around a different core (eg: a Power4 core + SIMD, or maybe 2) to give a very nice chip. This would also considerably speed up the development of the chip, any thoughts.
michael
<strong>Apple's G5 may come from IBM but it may be totally irrelevant to IBM's plans.
Given that IBM really does well in the server market and they are pressing ahead with Linux for everything they sell, it makes sense that they are targeting this chip family towards mainstream low-end serving duties. In the server space Linux does really well with a signicant and growing share of the total market which is far larger than the small declining desktop market Apple has. </strong><hr></blockquote>
Actually, this isn't true. Linux's penetration into the server space is actually on the decline -- at least temporarily -- and its desktop installations aren't even on the radar -- Apple came in with Mac OS X, and almost overnight surprassed Linux in terms of installed seats.
Linux for geeks and sysadmins to run on their servers is one thing; running it on the desktop is something it has failed miserably at, and rightly so: it sucks as a desktop OS. Do you honestly believe that there is a larger market for servers than there is for desktop computers? What would they be serving to, if that were the case?
[ 08-17-2002: Message edited by: moki ]</p>
[quote] Do you honestly believe that there is a larger market for servers than there is for desktop computers? What would they be serving to, if that were the case?<hr></blockquote>
What, you never heard of "self-serve?"
I think it's clear that these processors are aimed at Apple's desktop market. IBM already deploys Linux on commodity x86 platforms; what's the point of developing an entire new architecture to run the same old stuff?
If it ain't broke, don't fix it, so the saying goes.
Now Apple's pro processor situation- that's broke. I think IBM's fixin' to change it.
eg.
I thought I'd try the dual giggers again. Running Photoshop. Maybe it was just me last time. Maybe they were really fast.
Tried throwing around 8 by 11 inch 300 dpi images in PS7 on 'X'.
512 megs of ram. No small amount.
My opinion?
Pauseworthy. PS7 on 'X' drags its feet. The dual gigger crawls through many of the Photoshop filters. Not enough that you'd go away and make a cup of tea...but yeesh. This is Apple's flagship? (Well, it was 8 months ago...and still is...seeing as the tiny bump cheese grater won't be shipping until end of September.)
Not impressed.
My point? My wife's 384 megs/600 G3 iBook running Photoshop 4 on Classic doesn't suffer in a perceptional comparison. You may snigger at that...but...I'm not.
Apple are innovating with OS, Software and design.
Everything but the actual specs.
I don't think we're going to see the kind of performance to run X or its applications properly until we leave the G4 behind. It just can't hack it. It was designed to compete with low mhz Pentium 3...which are now part of Intel's budget range.
Somebody calls the dual giggers 'Slow Silvers' on these boards. He's right, y'know.
Having said that. I've recently got my wife's iBook onto a DSL connection and I haven't been able to stop using it for the internet! 'X' is great. 'Anti Aliased' type...gel buttons on Yahoo pages...its...beautiful. I've got it mirrored through a 21inch monitor (which used to 'mirror' my PowerMac 200mhz clone...) and 'X' looks beautiful. It's a work of art.
My real point?
I'll concede that you don't have to have a 'G5' to enjoy 'X' now...but to do everything Apple is trying to do now...JUSTICE...then we're going to need that G5.
It's the last piece of the puzzle.
I believe now that the final piece will come from IBM.
Lemon Bon Bon
Enjoying his wife's iBook. Still 'power'Mac less though. Come on Apple, release the darn G5 already...
Given that IBM really does well in the server market and they are pressing ahead with Linux for everything they sell, it makes sense that they are targeting this chip family towards mainstream low-end serving duties. In the server space Linux does really well with a significant and growing share of the total market which is far larger than the small declining desktop market Apple has. ]]]
This is totally incorrect. Forget market share. Look at the numbers of actual "users". Apple's user base is GROWING. So stop trolling.
IBM is NOT going to try to bring LINUX into the consumer mass-market. OS X is more developed, usable and more robust and ALREADY in the mass-market for consumers.
As a matter of FACT, Apple is now the LARGEST distributor of a commercial-grad UNIX-based operating system.
David K. Every said it best...
[[[OSX is already the unit leader in terms of a UNIX distribution, BUT... Other people that rely on UNIX (corporate entities et. al.) are learning that IT IS ALREADY THE INTERFACE LEADER as well. And this is 12 months out of the gate. Jaguar looks to fix/improve many things... And in another 12-18 months, I think it will be the standard by which most UNIXXES are measured... OSX delivers today on what LINUX has been promising for 10 years... ]]] - DKE
IBM is obviously aware of this as well. Now how does that figure into the equation and Why this isn't a signal that IBM is going to try another attempt to win-over the desktop? Well, all you have to do is think about it...
What reason would IBM all of a sudden start to utilize AltiVec (or some other version of the original VMX design)? The embedded market? Unlikely, IBM stated that they developed a CPU *for the desltop*. Now, why would they do that? What other OS runs on PPC, LINUX? OK, I'll buy that.. But what would LINUX need AltiVec for? OK, maybe there are a *few* obscure, not-often-used-by-average-people apps. Still, there are NONE that are used in the mainstream.
Would these machines be running a PPC-LINUX version of Photoshop with AltiVec optimizations? No.
What about a LINUX version of Lightwave or Maya with AltiVec optimizations? Nope.
Office and AutoCAD for LINUX? No.
Will there suddenly be a massive rush of developers to code any of this stuff for IBM? I doubt it.
That leads to only one, logical conclusion...
Apple *already* has the OS that utilizes both the PPC and AltiVec quite well. There are PLENTY of apps already in place that have been utilizing these technologies for a while now and the brand is well-known. For developers, It is cheaper to make it work on OSX than other UNIXXEN, and OS X already has a larger installed base.
And as for people who think that the vector unit on the IBM cpu *isn't* AltiVec, let me remind you that AltiVec is just as much a technology of IBM as it is to Apple and Motorola. IBM wouldn't need special licensing. Furthermore, It's unlikely to believe that IBM would go through all the trouble to develop a completely new vector unit when they have a perfectly good one already available. Additionally, why would they bother to develop new tools to work with the new vector unit and how much training would be involved trying to get developers accustomed to using the technology properly?
Apple has the OS in place AND the commercial apps that LINUX simply doesn't have. OS X has a UI that's much further along and continuing to make strides.
--
Ed M.