I'm amazed no one realizes one of the biggest gems in Apple's crown is not just OS X, but Cocoa. It's one helluva powerful AppKit and I believe Apple is making a ton of headway by finally including Developer Tools with every machine sold and allowing free access to it on the Net.
A lot of people stick with Windows because it was the easiest to rapidly develop in house apps on (NeXTStep not included) with VisualBASIC and Visual C++.
Now, however, the shoe is on the other foot. MSFT is charging big dollars for dev tools and Apple is giving away very advanced development kits for free, and even better its on a UNIX system. And with the ability to use terminal commands inside a Cocoa app, it makes giving user unfriendly command line utils an easy to build front end.
I dare say that the geeks will be getting stoked as they learn more about Cocoa. And, well, where the programmers are is were the programs are.
The 970 will be huge and help draw more folks into the Mac, but Cocoa will help keep them here with a wider variety of apps for the platform.
<strong>I'm amazed no one realizes one of the biggest gems in Apple's crown is not just OS X, but Cocoa. It's one helluva powerful AppKit and I believe Apple is making a ton of headway by finally including Developer Tools with every machine sold and allowing free access to it on the Net.
A lot of people stick with Windows because it was the easiest to rapidly develop in house apps on (NeXTStep not included) with VisualBASIC and Visual C++.
Now, however, the shoe is on the other foot. MSFT is charging big dollars for dev tools and Apple is giving away very advanced development kits for free, and even better its on a UNIX system. And with the ability to use terminal commands inside a Cocoa app, it makes giving user unfriendly command line utils an easy to build front end.
I dare say that the geeks will be getting stoked as they learn more about Cocoa. And, well, where the programmers are is were the programs are.
The 970 will be huge and help draw more folks into the Mac, but Cocoa will help keep them here with a wider variety of apps for the platform.</strong><hr></blockquote>
The big problem with Cocoa is that it is Object C. Which is not in itself the problem, but the vast majoriaty of progrmmers do not know this varient of the C programming language.
<strong>The big problem with Cocoa is that it is Object C. Which is not in itself the problem, but the vast majoriaty of progrmmers do not know this varient of the C programming language.</strong><hr></blockquote>
Its not that much of a problem, actually. Obj-C is a very simple extension to C and Cocoa is the only framework I've used so far (out of about half a dozen) in which I've managed to build a reasonably complex application in mere days... without having to fight with the framework to make it happen. Instead I can just deal with the parts of my app which make it my app. I must say that I'm quite impressed. I haven't tried it but apparently Java can be used in place of Obj-C if that's more to your liking (its not to mine).
Apple can't rest on its laurels, however, because Microsoft's .NET environments (MC++, C#, etc) are quite impressive as well. And Apple's development tools have some stability issues which I'm currently struggling with (of course, so does .NET!).
Its not that much of a problem, actually. Obj-C is a very simple extension to C and Cocoa is the only framework I've used so far (out of about half a dozen) in which I've managed to build a reasonably complex application in mere days... without having to fight with the framework to make it happen. Instead I can just deal with the parts of my app which make it my app. I must say that I'm quite impressed. I haven't tried it but apparently Java can be used in place of Obj-C if that's more to your liking (its not to mine).
.</strong><hr></blockquote>
I'd have to completely agree with Programmer - my day job is Carbon, and I've been playing with Cocoa, and doing some UI prototyping for usability testing with it, and I was amazed at how quickly I got something really nice up and running.
Others have told me Cocoa gets in the way with large apps, but those who have told me this are old, old Carbon & framework folks, who have no experience with Cocoa. Judging by apps like iPhoto and iDVD though, I think it's like anything, the more experience you have with a framework, the better app you can write.
no one seems to have tackled one issue apple must be concerned about if/when they release a 970 based powermac.
the large jump in clock speed and performance will take its toll on the perceived power of apple's other products. true, the xserve may go 970 based as new powermacs are rolled out, but with the reported Mhz limitations on the current iteration of the G4, won't there be a tremendous disparity between powermacs and portables and consumer machines?
it is quite a leap of faith to think that apple would rev their whole product line simulatiously in order to address this. moving emacs and imacs to 1/1.25 or 1.42 Ghz G4s, a 1Ghz+ G3 or even a G4 for the iBook, and some serious Mhz leaps for the current 3 Powerbooks would seem like the most likely option. even if the 970 is intro'd at a max 1.8Ghz, those other machines will look increasing beleagured in terms of price, especially to consumers comparing them to the lastest x86 machines on the shelves, never mind the new cool looking pro macs.
the original post didn't seem to propose this as one reason apple may be in a pricing and perception quandary in the future. how apple promotes and prices the new 970 based powermacs will have a tremendous impact on their bottom line. i'm sure that, if apple can deliver the new machines on time and without too many migratory kludges, they will sell like hotcakes. if they complement g4 powermacs in the product range they can be priced higher, but will still sell to many pros if they are fast and at the bleeding edge.
however, apple will have difficulty maintaining current price points on their other products. surely they must be aware of this potential problem. how will they deal with it? pin their hopes on the power of 970s and X 10.3, and strong new sales in core and developing markets? promote an aggressive hardware business more focused on the digital hub, and thus a potential lucrative and non Mhz dependent revenue stream? run consumer and laptop prices down to an even tighter set of profit margins to try and stimulate the mac market share, whilst making their machines look cheaper compared to x86s?
personally, i'm stuck on how to get around it. i don't doubt that the 970 is coming to macs this year. i just don't know how they will keep it from affecting the rest of the company's hardware products, without some other 'magic' restorative component. mebbe a new killer app or device. a new hitherto unknown processor or scaling technique to allow other machines to keep pace in the 970's slipstream more gracefully. or something we really don't know about ... a strategic alliance with another company, or some revolutionary new way of using a mac. answers on a postcard please ....
[quote]Earlier in the month, I reported on the possibility that the PPC 970 might be released at 2.5GHz. This was big news, but it seems like most of us might have been a little too optimistic. I just ran across a link at Ace's to an older ZDNET article with the following tidbit in it:
The company first revealed its plans for the chip at last October's Microprocessor Forum, and at the time said the chip would run up to 1.8GHz, suggesting that a manufacturing process upgrade would be necessary before higher clock speeds could be achieved. IBM says a later PowerPC 970 will reach 2.5GHz using a 90-nanometre manufacturing process.
The first run will be out of IBM's 0.13u SOI Fishkill, NY fab, so it looks like we won't see the 2.5GHz version as soon as we'd hoped. (Also, while you're flaming me about the Salon post below, be sure and flame me about Part II of the PPC 970 article. My Inbox has been a little empty lately, and I'm beginning to feel unloved.)<hr></blockquote>
I think that Apples portables are in good shape. They will probably get another revision in late spring to early summer. Right now they are Apple most competative products in price performance. Also Laptops tend to lag behid in top end speed from their desktop cousins.
The big question I have is in the consumer market. Apple went 13 months last year without a speed bump, and when it came they were pretty conservative with it. The certainly had room to push the high end at least up to 1.25 Ghz, but didnt. Now I know all the arguments about what is really needed for home use and all, but the iMacs are too slow to truely be competative in their market. Everyone likes car analogies (Not! <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> ), so to put it another way, Apple is trying to market a Ferrarie with an Escort engine in it with the current crop of iMacs, and it dosnt look like the new G4's will come along in time, or to go to Dual's. The only way out that I see is for them to move them to 970's as well. If rumour of 2 new mother boards bieng ready for production holds true, I would think that these are for the PM and Xserve, and we will see another lackluster iMac speed bump in late summer to early fall.
I wouldn't fret about it -- this looks like its based on a mix of new and old information, all of it supposition. The 0.13 micron version can probably reach 2.5 GHz, it will just generate a whole lot of heat (65 W according to the French site). I doubt Apple will use such a hot processor in an MP machine.
>however, apple will have difficulty maintaining current >price points on their other products. surely they must >be aware of this potential problem. how will they deal >with it? pin their hopes on the power of 970s and X >10.3, and strong new sales in core and developing >markets? promote an aggressive hardware business >more focused on the digital hub, and thus a potential >lucrative and non Mhz dependent revenue stream? >run consumer and laptop prices down to an even >tighter set of profit margins to try and stimulate the mac >market share, whilst making their machines look >cheaper compared to x86s?
No one seems to address one issue which ought to be obvious. The components which Apple uses in its hardware cost them about the same as they cost every other manufacturer subject to volume pricing differentials (one advantage of selling more machines).
Regarding the upcoming performance gap between the 970 and non 970 iSomething computers. With 970 in Towers ans Servers the performence gap will get begger than today, but: First of all there alredy is a big difference one 1 GH G4 in top of the line iMac and iBooks compared to the duals in the Towers. Secondly if you really want performence in the first place you do not buy a iMac.
The power requirment of the first generation of 970 at 1.2 GHz are compatible with both mobile use and things like the iMac. So my guess for 970 inplementation order is 1. Tower 2. Server. 3. PBook. 4 iMac. 5. iBook. 1&2 will com at the same time more or less, 3 not that far behind and the other at a later time 2 or 3 upgrade cycles behind the start. (all this assuming that there are no bus or memory suff that is problematic in the confinements of a BP or iMac)
AppleNut!
You do not have to don the asbestos shirt due to me flaming you, but I am really eager to read that part II about 970
All of the sudden, I've got the urge for a little EV Nova
Heh, heh! Just fired up EVN again last night. Still trying to figure out the Auroran thread.
Way to say it Moki. Apple has taken Motorola's mediocre production and wrung every ounce of performance out of it. They have NOT held back. There is no reason to think they will start now.
If you want to look at theoretical numbers, the P4 3.06 should be more then twice as fast as a single 1.42 G4. In the real world while it is a lot faster, it isn't even close to twice as fast. (and we are not talking Altivec or multiprocessor here). Apple has done some amazing things with what they had to work with. Now they have something more astounding to work with and there is not a single indication in their history under Jobs that indicates they will hold anything back on the top line PowerMacs. I fully expect them to wring every ounce of performance out of the 970 that they can, and to use the fastest chips that IBM can provide them in sufficient volume.
As to the rumored 2.5ghz? It was a chip in a lab that demonstrated what the current process was capable of on the high end, not what is expected of current volume production. A dual 1.8ghz PowerMac will be a heck of a machine but some folks wouldn't be satisfied if it was a quad 5ghz. They would probably still insist that Apple was holding back.
No one seems to have tackled one issue Apple must be concerned about if/when they release a 970 based powermac.
The large jump in clock speed and performance will take its toll on the perceived power of apple's other products. true, the xserve may go 970 based as new powermacs are rolled out, but with the reported Mhz limitations on the current iteration of the G4, won't there be a tremendous disparity between powermacs and portables and consumer machines?
Not if they adjust the prices accordingly. The problem now is justifying a $1,499 + $799 Power Mac and 17" Studio Display over a $1799 iMac. If the 970 Power Macs force a lower iMac price, then it's a good thing.
I curious about how often Apple will update its powermacs now that Moto isn't doling out the speed bumps. If Apple has to wait long between 970 speed bumps, it would make sense for them to bump Powermacs more often to maximize profits. Something like this:
IBM's 970 (top end):
1.8 GHz June 03
2.5 GHz June 04
3.4 GHz June 05
Powermacs (top end):
1.4 GHz June 03
1.8 GHz Jan 04
2.2 GHz June 04
2.5 GHz Jan 05
3.0 GHz June 05
3.4 GHz Jan 06
Of course this makes sense for maximizing profits only if IBM bumps the 970 at long intervals. If IBM is giving Apple speed bumped 970s every 4 months, it wouldn't make sense for Apple to bump the towers every two months, the cost of manufacturing switchover and inventory management would be too high. My question is, how often will Apple bump the Powermacs? If Apple has the option to bump them every 5 months, will they? Every 6 months? And how often will IBM bump the 970? If IBM is bumping it every 6 months, then Apple won't have much of a choice in the matter, they've got to offer the best they have, since their only alternative would be to speedbump the towers every 3 months.
Yes, there is of course the need to offer very conservatively bumped Powermacs to maintain a buffer against IBM hitting their own speedbumps in scaling the 970. But with the 970, IBM has an interest in scaling it for themselves, since they will use it in blade servers. Self-interest, along with a killer reputation at crafting silicon, should render this "moto-buffer" strategy obsolete.
Besides GHz, Apple could hold back on dual CPUs...this is more probably but still unlikely. I like to think that Apple knows about the roaring inferno under their ass, and is prepared to offer the fastest Powermacs possible.
I curious about how often Apple will update its powermacs now that Moto isn't doling out the speed bumps. If Apple has to wait long between 970 speed bumps, it would make sense for them to bump Powermacs more often to maximize profits. Something like this:
JD,
I think one way of looking at how fast they are going to roll them out is look at the present PMs. Apple is using the fastest chips that are available. At least according to Motorola's website. Nowhere on their website (that I am aware of) does it discuss 1.42 GHz G4s, yet Apple is using them. I think what Moki discussed above is correct. Apple knows they are behind and is doing everything they can to catch up. They will release the fastest 970s that IBM can supply.
The premise that there is no competition is too flawed for words. If there was no competition for OSX, Apple's market share would be steadily climbing despite whatever insane price tag they put on their machines. Since they have actually lost more marketshare since Jobs' return, it must be the case (as it was before) that for the overwhelming majority of people windows offerings are not only competitive, they are superior.
People buy a machine (not an OS), this is a balance of PRICE, PERFORMANCE, and EXPERIENCE. Mebbe Apple provides a better experience, but it gets so soundly thrashed in the areas of price and performance, that it's still losing quarterly sales (especially in the pro line). The iBook has been one of their best selling machines. Why? because it's the closest they come to providing a cheap machine with relatively competitive performance. That means that OSX has LOTS of COMPETITION whether or not Apple chooses to recognize it or stick their faces in the sand for another 4 years.
This is the last chance Apple will see to make any serious inroads in their market-share -- which absolutely MUST INCREASE if Apple wants to remain a viable platform for another decade -- with the wintelon world's 64 bit path far from clear, and easily 18-24 months from offering the kind of reach a PPC970 might offer, this is the time to strike.
Not only do the machines need to be the fastest they can possibly be, they also need to, at a minimum, begin to drive their entire desktop line (consumer and pro) from 200-500USD cheaper on each and ever model.
If they choose to eek out the speed gains, people will hold off, the machines need to be basically irresistable. Jobs needs to get on stage and run the baddest photoshop/video/3-d bake off you've ever seen, something against a top shelf PC, not a carefully hobbled one, and when he does it, the mac has to win by at least half, but more like two thirds or three quarters. What takes 2 minutes on the mac, has to take 5 or more on the PC. Apple won't get there by holding back. Take stock of just how many pro customers they've lost, PM sales are a joke, and those customers aren't coming back unless they see dramatic performance increases. Remember they're now on a different platform, getting them to switch back after they've shelled out for PC systems and software won't be easy for the same reasons that many stuck with the mac for so long in the first place, switching platforms costs money and simply isn't worth it just to be a bit faster, you've gotta be 200-400% faster at the prices Apple is looking to charge. Or you gotta be a much more competitively priced! simple.
I hope Apple remembers this as the 970 hits and tumbles into the consumer range come 2004.
If it becomes that pervasive...it should be much cheaper than the G4 is now for Apple. That means lower prices AND profits. Except bigger profits. In that context and for the foreseeable future...single cpu supplier = This is a Good ThingTM.
Why? Because they'll be selling more of everything.
Sure... I think the sentiments you express have no merit. Apple has been providing the fastest machines it has been capable of providing; they have not tried to hold anything back, they have been hamstrung by a chip supplier (Motorola) that has lost interest in the desktop market.
With respect I think you're wrong. I think Apple are in the business of holding back - whether it be due to neccesity or profit. Their primary concern is their share price, and giving away the biggest break they've had in years isn't going to win them any friends on the stockmarket.
The evidence:
1. The illusive 1GHz G3 part, that's been mentioned on these very boards on countless occasions. This part is STILL not used in any Apple product - that's what I would call holding back! It would have been in the iBook by now if Apple weren't trying to make the Power Book G4 look more powerful by comparison.
2. Power Mac G4 line up is currrently 867MHz/1.25GHz/1.42GHz. What happened to the 933MHz & 1.0GHz parts? Did Motorola lose the plans overnight? The 867MHz part debuted in the initial Quicksilver G4 for fecks sake! If apple were in the business of delivering the fastest parts they could lay their hands on, the Power Mac G4 family would be at 1GHz/1.25GHz/1.42GHz. Again, they're not putting the 933MHz or 1GHz parts into the entry level machine because the two other machines need to look fast by comparison.
I can guarantee, that they'll be looking at the parts that are available to them, and figuring out a way of squeezing every last penny out of this new CPU. Apple have always done this, it means you can sell the same machine to three different customers at the highest price that each of those customers is willing to pay. You don't just sell it at the cheapest price! Look at all the money you'd lose!
The frequencies themselves don't mean anything. I'm not about to predict the exact frequencies that we'll see. What's important is that the new machines must deliver more bang per buck than anything that has gone before. I can guarantee that if they released a 1.0GHz/1.2GHz/1.4GHz G5 family it would sell, because each of those machines is considerably faster than the machine it replaces.
What they won't do is release the fastest parts available to them on day one, and then sit on their hands and hope blindly that IBM can deliver on it's promise.
Just because a 2.5GHz part is available, we won't see a 2.3GHz/2.4GHz/2.5GHz line up. We might see a 1.4GHz/2.0GHz/2.4GHz line up, with Apple having all the freqencies in between as options for the future.
They might CLAIM that they are the fastest parts available, but they'll play it smart, they'll dripfeed us the parts and make extra profits on the speed bumps, whilst stockpiling faster parts so that they can plan ahead. It's good business sense, and ergo: we'll end up paying top dollar for whatever they produce on the day, only to have them wave it under our noses at a vastly reduced price shortly thereafter.
Just because you know more about the Mac than most people, and want the fastest processor available, isn't enough to actually make it happen.
Can you provide evidence that it exists? I can't find any devices that use a 1GHz G3, but then, I didn't look very hard, either. I'm not so sure the chip exists, yet.
Quote:
2. Power Mac G4 line up is currrently 867MHz/1.25GHz/1.42GHz.[...]
As Flounder pointed out, you're quite simply wrong about this one.
Even if the above points were true, you'd still have a pretty big hole in your argument. The above examples show that Apple is willing to hold back the low end of their line in order to make the high end look better... That's a far cry from holding back the high end to begin with.
Comments
<strong>
I'm with Moki. Sooner or later, Apple will stand up to the plate and...STRIKE.
</strong><hr></blockquote>
Whoa... Think you meant "out of the park grand slam."
A lot of people stick with Windows because it was the easiest to rapidly develop in house apps on (NeXTStep not included) with VisualBASIC and Visual C++.
Now, however, the shoe is on the other foot. MSFT is charging big dollars for dev tools and Apple is giving away very advanced development kits for free, and even better its on a UNIX system. And with the ability to use terminal commands inside a Cocoa app, it makes giving user unfriendly command line utils an easy to build front end.
I dare say that the geeks will be getting stoked as they learn more about Cocoa. And, well, where the programmers are is were the programs are.
The 970 will be huge and help draw more folks into the Mac, but Cocoa will help keep them here with a wider variety of apps for the platform.
<strong>I'm amazed no one realizes one of the biggest gems in Apple's crown is not just OS X, but Cocoa. It's one helluva powerful AppKit and I believe Apple is making a ton of headway by finally including Developer Tools with every machine sold and allowing free access to it on the Net.
A lot of people stick with Windows because it was the easiest to rapidly develop in house apps on (NeXTStep not included) with VisualBASIC and Visual C++.
Now, however, the shoe is on the other foot. MSFT is charging big dollars for dev tools and Apple is giving away very advanced development kits for free, and even better its on a UNIX system. And with the ability to use terminal commands inside a Cocoa app, it makes giving user unfriendly command line utils an easy to build front end.
I dare say that the geeks will be getting stoked as they learn more about Cocoa. And, well, where the programmers are is were the programs are.
The 970 will be huge and help draw more folks into the Mac, but Cocoa will help keep them here with a wider variety of apps for the platform.</strong><hr></blockquote>
The big problem with Cocoa is that it is Object C. Which is not in itself the problem, but the vast majoriaty of progrmmers do not know this varient of the C programming language.
<strong>The big problem with Cocoa is that it is Object C. Which is not in itself the problem, but the vast majoriaty of progrmmers do not know this varient of the C programming language.</strong><hr></blockquote>
Its not that much of a problem, actually. Obj-C is a very simple extension to C and Cocoa is the only framework I've used so far (out of about half a dozen) in which I've managed to build a reasonably complex application in mere days... without having to fight with the framework to make it happen. Instead I can just deal with the parts of my app which make it my app. I must say that I'm quite impressed. I haven't tried it but apparently Java can be used in place of Obj-C if that's more to your liking (its not to mine).
Apple can't rest on its laurels, however, because Microsoft's .NET environments (MC++, C#, etc) are quite impressive as well. And Apple's development tools have some stability issues which I'm currently struggling with (of course, so does .NET!).
<strong>
Its not that much of a problem, actually. Obj-C is a very simple extension to C and Cocoa is the only framework I've used so far (out of about half a dozen) in which I've managed to build a reasonably complex application in mere days... without having to fight with the framework to make it happen. Instead I can just deal with the parts of my app which make it my app. I must say that I'm quite impressed. I haven't tried it but apparently Java can be used in place of Obj-C if that's more to your liking (its not to mine).
.</strong><hr></blockquote>
I'd have to completely agree with Programmer - my day job is Carbon, and I've been playing with Cocoa, and doing some UI prototyping for usability testing with it, and I was amazed at how quickly I got something really nice up and running.
Others have told me Cocoa gets in the way with large apps, but those who have told me this are old, old Carbon & framework folks, who have no experience with Cocoa. Judging by apps like iPhoto and iDVD though, I think it's like anything, the more experience you have with a framework, the better app you can write.
the large jump in clock speed and performance will take its toll on the perceived power of apple's other products. true, the xserve may go 970 based as new powermacs are rolled out, but with the reported Mhz limitations on the current iteration of the G4, won't there be a tremendous disparity between powermacs and portables and consumer machines?
it is quite a leap of faith to think that apple would rev their whole product line simulatiously in order to address this. moving emacs and imacs to 1/1.25 or 1.42 Ghz G4s, a 1Ghz+ G3 or even a G4 for the iBook, and some serious Mhz leaps for the current 3 Powerbooks would seem like the most likely option. even if the 970 is intro'd at a max 1.8Ghz, those other machines will look increasing beleagured in terms of price, especially to consumers comparing them to the lastest x86 machines on the shelves, never mind the new cool looking pro macs.
the original post didn't seem to propose this as one reason apple may be in a pricing and perception quandary in the future. how apple promotes and prices the new 970 based powermacs will have a tremendous impact on their bottom line. i'm sure that, if apple can deliver the new machines on time and without too many migratory kludges, they will sell like hotcakes. if they complement g4 powermacs in the product range they can be priced higher, but will still sell to many pros if they are fast and at the bleeding edge.
however, apple will have difficulty maintaining current price points on their other products. surely they must be aware of this potential problem. how will they deal with it? pin their hopes on the power of 970s and X 10.3, and strong new sales in core and developing markets? promote an aggressive hardware business more focused on the digital hub, and thus a potential lucrative and non Mhz dependent revenue stream? run consumer and laptop prices down to an even tighter set of profit margins to try and stimulate the mac market share, whilst making their machines look cheaper compared to x86s?
personally, i'm stuck on how to get around it. i don't doubt that the 970 is coming to macs this year. i just don't know how they will keep it from affecting the rest of the company's hardware products, without some other 'magic' restorative component. mebbe a new killer app or device. a new hitherto unknown processor or scaling technique to allow other machines to keep pace in the 970's slipstream more gracefully. or something we really don't know about ... a strategic alliance with another company, or some revolutionary new way of using a mac. answers on a postcard please ....
The company first revealed its plans for the chip at last October's Microprocessor Forum, and at the time said the chip would run up to 1.8GHz, suggesting that a manufacturing process upgrade would be necessary before higher clock speeds could be achieved. IBM says a later PowerPC 970 will reach 2.5GHz using a 90-nanometre manufacturing process.
The first run will be out of IBM's 0.13u SOI Fishkill, NY fab, so it looks like we won't see the 2.5GHz version as soon as we'd hoped. (Also, while you're flaming me about the Salon post below, be sure and flame me about Part II of the PPC 970 article. My Inbox has been a little empty lately, and I'm beginning to feel unloved.)<hr></blockquote>
uh oh
<a href="http://www.arstechnica.com"; target="_blank">www.arstechnica.com</a>
I think that Apples portables are in good shape. They will probably get another revision in late spring to early summer. Right now they are Apple most competative products in price performance. Also Laptops tend to lag behid in top end speed from their desktop cousins.
The big question I have is in the consumer market. Apple went 13 months last year without a speed bump, and when it came they were pretty conservative with it. The certainly had room to push the high end at least up to 1.25 Ghz, but didnt. Now I know all the arguments about what is really needed for home use and all, but the iMacs are too slow to truely be competative in their market. Everyone likes car analogies (Not! <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> ), so to put it another way, Apple is trying to market a Ferrarie with an Escort engine in it with the current crop of iMacs, and it dosnt look like the new G4's will come along in time, or to go to Dual's. The only way out that I see is for them to move them to 970's as well. If rumour of 2 new mother boards bieng ready for production holds true, I would think that these are for the PM and Xserve, and we will see another lackluster iMac speed bump in late summer to early fall.
<strong>
uh oh
<a href="http://www.arstechnica.com"; target="_blank">www.arstechnica.com</a></strong><hr></blockquote>
I wouldn't fret about it -- this looks like its based on a mix of new and old information, all of it supposition. The 0.13 micron version can probably reach 2.5 GHz, it will just generate a whole lot of heat (65 W according to the French site). I doubt Apple will use such a hot processor in an MP machine.
No one seems to address one issue which ought to be obvious. The components which Apple uses in its hardware cost them about the same as they cost every other manufacturer subject to volume pricing differentials (one advantage of selling more machines).
[ 03-10-2003: Message edited by: RBR ]</p>
The power requirment of the first generation of 970 at 1.2 GHz are compatible with both mobile use and things like the iMac. So my guess for 970 inplementation order is 1. Tower 2. Server. 3. PBook. 4 iMac. 5. iBook. 1&2 will com at the same time more or less, 3 not that far behind and the other at a later time 2 or 3 upgrade cycles behind the start. (all this assuming that there are no bus or memory suff that is problematic in the confinements of a BP or iMac)
AppleNut!
You do not have to don the asbestos shirt due to me flaming you, but I am really eager to read that part II about 970
Originally posted by Flounder
Moki, laying on the smack down.
All of the sudden, I've got the urge for a little EV Nova
Heh, heh! Just fired up EVN again last night. Still trying to figure out the Auroran thread.
Way to say it Moki. Apple has taken Motorola's mediocre production and wrung every ounce of performance out of it. They have NOT held back. There is no reason to think they will start now.
If you want to look at theoretical numbers, the P4 3.06 should be more then twice as fast as a single 1.42 G4. In the real world while it is a lot faster, it isn't even close to twice as fast. (and we are not talking Altivec or multiprocessor here). Apple has done some amazing things with what they had to work with. Now they have something more astounding to work with and there is not a single indication in their history under Jobs that indicates they will hold anything back on the top line PowerMacs. I fully expect them to wring every ounce of performance out of the 970 that they can, and to use the fastest chips that IBM can provide them in sufficient volume.
As to the rumored 2.5ghz? It was a chip in a lab that demonstrated what the current process was capable of on the high end, not what is expected of current volume production. A dual 1.8ghz PowerMac will be a heck of a machine but some folks wouldn't be satisfied if it was a quad 5ghz. They would probably still insist that Apple was holding back.
Originally posted by jobes
No one seems to have tackled one issue Apple must be concerned about if/when they release a 970 based powermac.
The large jump in clock speed and performance will take its toll on the perceived power of apple's other products. true, the xserve may go 970 based as new powermacs are rolled out, but with the reported Mhz limitations on the current iteration of the G4, won't there be a tremendous disparity between powermacs and portables and consumer machines?
Not if they adjust the prices accordingly. The problem now is justifying a $1,499 + $799 Power Mac and 17" Studio Display over a $1799 iMac. If the 970 Power Macs force a lower iMac price, then it's a good thing.
IBM's 970 (top end):
1.8 GHz June 03
2.5 GHz June 04
3.4 GHz June 05
Powermacs (top end):
1.4 GHz June 03
1.8 GHz Jan 04
2.2 GHz June 04
2.5 GHz Jan 05
3.0 GHz June 05
3.4 GHz Jan 06
Of course this makes sense for maximizing profits only if IBM bumps the 970 at long intervals. If IBM is giving Apple speed bumped 970s every 4 months, it wouldn't make sense for Apple to bump the towers every two months, the cost of manufacturing switchover and inventory management would be too high. My question is, how often will Apple bump the Powermacs? If Apple has the option to bump them every 5 months, will they? Every 6 months? And how often will IBM bump the 970? If IBM is bumping it every 6 months, then Apple won't have much of a choice in the matter, they've got to offer the best they have, since their only alternative would be to speedbump the towers every 3 months.
Yes, there is of course the need to offer very conservatively bumped Powermacs to maintain a buffer against IBM hitting their own speedbumps in scaling the 970. But with the 970, IBM has an interest in scaling it for themselves, since they will use it in blade servers. Self-interest, along with a killer reputation at crafting silicon, should render this "moto-buffer" strategy obsolete.
Besides GHz, Apple could hold back on dual CPUs...this is more probably but still unlikely. I like to think that Apple knows about the roaring inferno under their ass, and is prepared to offer the fastest Powermacs possible.
Originally posted by Junkyard Dawg
I curious about how often Apple will update its powermacs now that Moto isn't doling out the speed bumps. If Apple has to wait long between 970 speed bumps, it would make sense for them to bump Powermacs more often to maximize profits. Something like this:
JD,
I think one way of looking at how fast they are going to roll them out is look at the present PMs. Apple is using the fastest chips that are available. At least according to Motorola's website. Nowhere on their website (that I am aware of) does it discuss 1.42 GHz G4s, yet Apple is using them. I think what Moki discussed above is correct. Apple knows they are behind and is doing everything they can to catch up. They will release the fastest 970s that IBM can supply.
The premise that there is no competition is too flawed for words. If there was no competition for OSX, Apple's market share would be steadily climbing despite whatever insane price tag they put on their machines. Since they have actually lost more marketshare since Jobs' return, it must be the case (as it was before) that for the overwhelming majority of people windows offerings are not only competitive, they are superior.
People buy a machine (not an OS), this is a balance of PRICE, PERFORMANCE, and EXPERIENCE. Mebbe Apple provides a better experience, but it gets so soundly thrashed in the areas of price and performance, that it's still losing quarterly sales (especially in the pro line). The iBook has been one of their best selling machines. Why? because it's the closest they come to providing a cheap machine with relatively competitive performance. That means that OSX has LOTS of COMPETITION whether or not Apple chooses to recognize it or stick their faces in the sand for another 4 years.
This is the last chance Apple will see to make any serious inroads in their market-share -- which absolutely MUST INCREASE if Apple wants to remain a viable platform for another decade -- with the wintelon world's 64 bit path far from clear, and easily 18-24 months from offering the kind of reach a PPC970 might offer, this is the time to strike.
Not only do the machines need to be the fastest they can possibly be, they also need to, at a minimum, begin to drive their entire desktop line (consumer and pro) from 200-500USD cheaper on each and ever model.
If they choose to eek out the speed gains, people will hold off, the machines need to be basically irresistable. Jobs needs to get on stage and run the baddest photoshop/video/3-d bake off you've ever seen, something against a top shelf PC, not a carefully hobbled one, and when he does it, the mac has to win by at least half, but more like two thirds or three quarters. What takes 2 minutes on the mac, has to take 5 or more on the PC. Apple won't get there by holding back. Take stock of just how many pro customers they've lost, PM sales are a joke, and those customers aren't coming back unless they see dramatic performance increases. Remember they're now on a different platform, getting them to switch back after they've shelled out for PC systems and software won't be easy for the same reasons that many stuck with the mac for so long in the first place, switching platforms costs money and simply isn't worth it just to be a bit faster, you've gotta be 200-400% faster at the prices Apple is looking to charge. Or you gotta be a much more competitively priced! simple.
I hope Apple remembers this as the 970 hits and tumbles into the consumer range come 2004.
If it becomes that pervasive...it should be much cheaper than the G4 is now for Apple. That means lower prices AND profits. Except bigger profits. In that context and for the foreseeable future...single cpu supplier = This is a Good ThingTM.
Why? Because they'll be selling more of everything.
Roll on 2004.
Lemon Bon Bon
Quote:
Sure... I think the sentiments you express have no merit. Apple has been providing the fastest machines it has been capable of providing; they have not tried to hold anything back, they have been hamstrung by a chip supplier (Motorola) that has lost interest in the desktop market.
With respect I think you're wrong. I think Apple are in the business of holding back - whether it be due to neccesity or profit. Their primary concern is their share price, and giving away the biggest break they've had in years isn't going to win them any friends on the stockmarket.
The evidence:
1. The illusive 1GHz G3 part, that's been mentioned on these very boards on countless occasions. This part is STILL not used in any Apple product - that's what I would call holding back! It would have been in the iBook by now if Apple weren't trying to make the Power Book G4 look more powerful by comparison.
2. Power Mac G4 line up is currrently 867MHz/1.25GHz/1.42GHz. What happened to the 933MHz & 1.0GHz parts? Did Motorola lose the plans overnight? The 867MHz part debuted in the initial Quicksilver G4 for fecks sake! If apple were in the business of delivering the fastest parts they could lay their hands on, the Power Mac G4 family would be at 1GHz/1.25GHz/1.42GHz. Again, they're not putting the 933MHz or 1GHz parts into the entry level machine because the two other machines need to look fast by comparison.
I can guarantee, that they'll be looking at the parts that are available to them, and figuring out a way of squeezing every last penny out of this new CPU. Apple have always done this, it means you can sell the same machine to three different customers at the highest price that each of those customers is willing to pay. You don't just sell it at the cheapest price! Look at all the money you'd lose!
The frequencies themselves don't mean anything. I'm not about to predict the exact frequencies that we'll see. What's important is that the new machines must deliver more bang per buck than anything that has gone before. I can guarantee that if they released a 1.0GHz/1.2GHz/1.4GHz G5 family it would sell, because each of those machines is considerably faster than the machine it replaces.
What they won't do is release the fastest parts available to them on day one, and then sit on their hands and hope blindly that IBM can deliver on it's promise.
Just because a 2.5GHz part is available, we won't see a 2.3GHz/2.4GHz/2.5GHz line up. We might see a 1.4GHz/2.0GHz/2.4GHz line up, with Apple having all the freqencies in between as options for the future.
They might CLAIM that they are the fastest parts available, but they'll play it smart, they'll dripfeed us the parts and make extra profits on the speed bumps, whilst stockpiling faster parts so that they can plan ahead. It's good business sense, and ergo: we'll end up paying top dollar for whatever they produce on the day, only to have them wave it under our noses at a vastly reduced price shortly thereafter.
Just because you know more about the Mac than most people, and want the fastest processor available, isn't enough to actually make it happen.
Time will tell!
Originally posted by Messiah
If apple were in the business of delivering the fastest parts they could lay their hands on, the Power Mac G4 family would be at 1GHz/1.25GHz/1.42GHz.
uhh, it is.
Originally posted by Junkyard Dawg
Carmack already has Q3 optimized for dual PPC. Bribe him to add a few altivec optimizations, stand back, and be amazed.
It would be something else to see hard-core gamers start buying Powermacs because they get 1.5-2X the framerates of Wintels!
Actually Quake 3 IS optimized for Altivec.
Steve needs to do a Photoshop bakeoff... If he doesnt, well, then PC speeds is still in the lead.
1. The illusive 1GHz G3 part,[...]
Can you provide evidence that it exists? I can't find any devices that use a 1GHz G3, but then, I didn't look very hard, either. I'm not so sure the chip exists, yet.
2. Power Mac G4 line up is currrently 867MHz/1.25GHz/1.42GHz.[...]
As Flounder pointed out, you're quite simply wrong about this one.
Even if the above points were true, you'd still have a pretty big hole in your argument. The above examples show that Apple is willing to hold back the low end of their line in order to make the high end look better... That's a far cry from holding back the high end to begin with.