It's all so simple IMHO..... Everyone and his sister KNOWS with metaphysical certainty that Apple will be incorporating IBM's PPC 970 into Powermacs. The only variable is WHEN ? To stem the tide of defections to the dark side Apple should IMMEDIATELY announce the date when said systems will be available....the reason is simple : if you're in the market for a Powermac and you're reading all this you may think it'll be introduced anytime between late summer and Feb 2004.... that may be too much uncertainty.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
970 at WWDC?
- Thread starter MacRumors
- Start date
- Sort by reaction score
We have estimates for the 1.8GHz PPC970. It is faster than Intel's flagship Itanium 2 processor (although not faster than the 32-bit Pentium 4; SPEC numbers across bitness are hard to compare, however). AMD's Hammer processor gets higher SPEC numbers in theory, but that's been delayed so much that if/when it finally does come out the 970 will be well beyond just 1.8GHz.
Given that SPEC numbers tend to scale directly with processor speed (assuming a constant processor design and scale), the 970 at 2.5GHz would literally blow any current or anticipated processor out of the water on SPEC scores.
The G4 scores, if I remember correctly, in the <300 range on both SPECint and SPECfp; the 970 (again, 64-bit vs 32-bit, but this comparison favors the 32-bit proc) is 1000-1200 in both (i forget the precise numbers).
In other words: one 970 should be able to do the CPU grunt work of 3-4 G4s.
Add to that the fact that the FSB is so much faster (and really that is where the 970 gets most of its comparitive speed, as the G4 is highly-FSB-bound), and one 970 could easily do the job of dual G4s with aplomb. On the other hand, the advantages of having dual procs in terms of responsiveness and such are hard to accomplish with a single proc no matter how fast it is.
some people will still buy when they can
This would hurt sales, but some organizations buy when they can. Education, for example, makes purchases when they have budget available. They generally can't carry budget over from one period to another (at least that was how it worked at my last job (university)).
If most Universities (like mine) had a budget that ended in the summer, they will buy machines irregardless of what will be available two months later.
It is beginning to look like apple needs to prove to Mac clients that there is, in fact, a road map for the desktop. If they don't show something big in the pipe, they may loose desktop space. The only thing worse than having sales postponed is having mac sales replaced by PCs. These PC slots generally don't come back to the Mac later on.
... just my 2 cents.
This would hurt sales, but some organizations buy when they can. Education, for example, makes purchases when they have budget available. They generally can't carry budget over from one period to another (at least that was how it worked at my last job (university)).
If most Universities (like mine) had a budget that ended in the summer, they will buy machines irregardless of what will be available two months later.
It is beginning to look like apple needs to prove to Mac clients that there is, in fact, a road map for the desktop. If they don't show something big in the pipe, they may loose desktop space. The only thing worse than having sales postponed is having mac sales replaced by PCs. These PC slots generally don't come back to the Mac later on.
... just my 2 cents.
Check your facts... and REALITY!
Stand corrected... I think, since I'm not quite sure what that meant. (64 bit arithmetic and logic instructions? Hmmmm...)
Believe it or not a LOT of apps should and will most likely stay 32-bit!! There are not that many apps that actually benefit from 64-bit, people!! Come on, you're doing the same damn thing the Wintellers are doing when it comes to M/GHz! Just cause there's a bigger number in there, everyone's "cool dude! you GOT to get one no matter what!" without even knowing what it means!
64-bit computing is great for applications that must (and CAN of course, which none can so far on the Mac) make use of the increased scalability and performance of a 64-bit OS (which we ALSO don't have). Most high-end computing needs are already met with 32-bit applications on a 32-bit OS, so let's keep it a tad more in perspective.
Most applications should even STAY 32-Bit! Any application that does not require 64-bit features should remain a 32-bit binary. 32-bit applications will run on both the 32-bit and 64-bit PPCs btw. Most 32-bit applications will perform better compiled as a 32-bit binary, because more of the application binary fits in the computer's cache. When a 32-bit application is recompiled for 64 bits, the 64-bit binary will typically be larger than its 32-bit binary. With a given cache size available on a system, performance may actually decline because of a greater number of cache misses when running the 64-bit binary........ I'm just going to assume this made sense to SOME of you, the other ones just don't talk about it as if 64-bit is a MUST before you do.
Close enough I guess... but just to make it a tad more specific: A 64-bit processor, with 64-bit registers and a 64-bit integer data path, (as opposed to a 32-bit proc, as you said, that has flat addressing of up to 2^32 32-bit bytes, or about 4GB of memory) has flat addressing for 2^64 64-bit bytes, or 18 billion GB (or 18 Exabytes) of memory... which should do for M$ Word I think. (wonder how much THAT memory upgrade is gonna cost in the Apple-Store 😀)
Yipee, huh?.... nope. At least not until there is the OS to go along with it and foremost the APPS, and really only the very high-end apps in the area of 3D, video/film (and I'm not talking about FCP! 🙄 ) and databases (servers!) are gonna even notice anything, since in a 64-bit CPU, integer arithmetic is 64 bits providing greater performance and precision. Since most compilers support 64-bit data types, even on 32-bit CPUs, the main benefit for integer arithmetic is more performance on larger data types. And larger isn't referring to your PowerPoint presentation...
A true 64-bit operating system uses the 64-bit addressing and arithmetic abilities of the CPU to provide more system resources to the apps. A 64-bit operating system would allow larger files and more files, user IDs, shared library segments, and other resources than a 32-bit operating system, which like I said hardly plays a part in >70% of users! But at least OS X would finally be FAST... too bad that it apparently will need a 64-bit version to get there. 😡
In OTHER WORDS, there's NO WAY we're going to see some sort of COMPLETE switch to the 970, that would be complete idiocy, especially in terms of cost. Therefore you're only going to see the VERY TOP of the line go that route, if even, and the servers! So you can stop the "it will hurt sales" talk now. It will only make sense to the FEWEST to actually get one at first! Which BTW is the same for the ITANIUMS and HAMMERS of this world!! Get real!
These puppies are gonna need a completely new I/O architecture (high speed everything!) to even make sense (and not be another half-assed attempt of Apple's to save some face by relying on the ol' reality-distortion-field!), and that ain't exactly FREE...
Dey is gonna be 'spensive!!! (relatively speaking) so don't plan on jumping on the first boxes, unless daddy's got the platinum layin' around unattended... 😉 😛
... should nonetheless shut a bunch WinTrolls up, but won't change the fact that the "lower" boxes are still left painfully slow. The road to being able to say "the Mac is (one of) the fastest machines!" without having to add a few "if"s or "but"s is gonna be loooooooong... nuff said.
Stand corrected... I think, since I'm not quite sure what that meant. (64 bit arithmetic and logic instructions? Hmmmm...)
Believe it or not a LOT of apps should and will most likely stay 32-bit!! There are not that many apps that actually benefit from 64-bit, people!! Come on, you're doing the same damn thing the Wintellers are doing when it comes to M/GHz! Just cause there's a bigger number in there, everyone's "cool dude! you GOT to get one no matter what!" without even knowing what it means!
64-bit computing is great for applications that must (and CAN of course, which none can so far on the Mac) make use of the increased scalability and performance of a 64-bit OS (which we ALSO don't have). Most high-end computing needs are already met with 32-bit applications on a 32-bit OS, so let's keep it a tad more in perspective.
Most applications should even STAY 32-Bit! Any application that does not require 64-bit features should remain a 32-bit binary. 32-bit applications will run on both the 32-bit and 64-bit PPCs btw. Most 32-bit applications will perform better compiled as a 32-bit binary, because more of the application binary fits in the computer's cache. When a 32-bit application is recompiled for 64 bits, the 64-bit binary will typically be larger than its 32-bit binary. With a given cache size available on a system, performance may actually decline because of a greater number of cache misses when running the 64-bit binary........ I'm just going to assume this made sense to SOME of you, the other ones just don't talk about it as if 64-bit is a MUST before you do.
Close enough I guess... but just to make it a tad more specific: A 64-bit processor, with 64-bit registers and a 64-bit integer data path, (as opposed to a 32-bit proc, as you said, that has flat addressing of up to 2^32 32-bit bytes, or about 4GB of memory) has flat addressing for 2^64 64-bit bytes, or 18 billion GB (or 18 Exabytes) of memory... which should do for M$ Word I think. (wonder how much THAT memory upgrade is gonna cost in the Apple-Store 😀)
Yipee, huh?.... nope. At least not until there is the OS to go along with it and foremost the APPS, and really only the very high-end apps in the area of 3D, video/film (and I'm not talking about FCP! 🙄 ) and databases (servers!) are gonna even notice anything, since in a 64-bit CPU, integer arithmetic is 64 bits providing greater performance and precision. Since most compilers support 64-bit data types, even on 32-bit CPUs, the main benefit for integer arithmetic is more performance on larger data types. And larger isn't referring to your PowerPoint presentation...
A true 64-bit operating system uses the 64-bit addressing and arithmetic abilities of the CPU to provide more system resources to the apps. A 64-bit operating system would allow larger files and more files, user IDs, shared library segments, and other resources than a 32-bit operating system, which like I said hardly plays a part in >70% of users! But at least OS X would finally be FAST... too bad that it apparently will need a 64-bit version to get there. 😡
In OTHER WORDS, there's NO WAY we're going to see some sort of COMPLETE switch to the 970, that would be complete idiocy, especially in terms of cost. Therefore you're only going to see the VERY TOP of the line go that route, if even, and the servers! So you can stop the "it will hurt sales" talk now. It will only make sense to the FEWEST to actually get one at first! Which BTW is the same for the ITANIUMS and HAMMERS of this world!! Get real!
These puppies are gonna need a completely new I/O architecture (high speed everything!) to even make sense (and not be another half-assed attempt of Apple's to save some face by relying on the ol' reality-distortion-field!), and that ain't exactly FREE...
Dey is gonna be 'spensive!!! (relatively speaking) so don't plan on jumping on the first boxes, unless daddy's got the platinum layin' around unattended... 😉 😛
... should nonetheless shut a bunch WinTrolls up, but won't change the fact that the "lower" boxes are still left painfully slow. The road to being able to say "the Mac is (one of) the fastest machines!" without having to add a few "if"s or "but"s is gonna be loooooooong... nuff said.
I don't think so. Any new processor has limitted production capacity at first relative to when it is going "full bore". While there are relatively fewer potential users of the XServe, they are much more likely to spend the extra bucks on a new processor/MB design. This is why new processors are generally released to the "higher end" machines before the "lower end" machines. Also, note that each motherboard takes design resources to get out the door: putting 970's on all machines from day one would both be silly and would take Apple precious months to coordinate, when they could be selling XServes based on the new design while the PMs are under design, and sell PMs with the new design while working on bumping up the iMacs, etc.
Just a possibility; store it away and don't be too upset if that's what happens.
Hardware Keynote
Having been to over a dozen WWDCs...
WWDC has only been used for announcing hardware once, the Lombard. It has been used many times for teasing the developers with bake-offs using future, possible hardware. (Anyone remember the tri-media chip?) Not always have these products shipped and the lead time has been as much as a year.
WWDC is usually a platform for announcing software. An official announcement and release of a Puma beta to developers is most likely.
The demoing of prerelease hardware at WWDC can range from private, invitation-only affairs to a bake-off for 5,000 of my closest friends.
Neither one would surprise me this year. A clue will be if the Hardware Keynote reappears this year. Usually held on Wednesday morning. Another clue will be if there's a session on 64bit Programming Directions. We'll now when Apple posts the sessions in a few weeks.
I'd suspect that the 64bit session is more likely with a lot of wink, wink, nudge, nudge. But I'd love to be surprised.
Sebastian
Having been to over a dozen WWDCs...
WWDC has only been used for announcing hardware once, the Lombard. It has been used many times for teasing the developers with bake-offs using future, possible hardware. (Anyone remember the tri-media chip?) Not always have these products shipped and the lead time has been as much as a year.
WWDC is usually a platform for announcing software. An official announcement and release of a Puma beta to developers is most likely.
The demoing of prerelease hardware at WWDC can range from private, invitation-only affairs to a bake-off for 5,000 of my closest friends.
Neither one would surprise me this year. A clue will be if the Hardware Keynote reappears this year. Usually held on Wednesday morning. Another clue will be if there's a session on 64bit Programming Directions. We'll now when Apple posts the sessions in a few weeks.
I'd suspect that the 64bit session is more likely with a lot of wink, wink, nudge, nudge. But I'd love to be surprised.
Sebastian
hee hee...
just found a cute comparison here, where they're talking about "A CPU's ability to address data depends on how many bits it can handle at one time":
8-bit
256
About 10 times larger than our alphabet
16-bit
65,536
Population of a midsize city
32-bit
4.3 billion
Total adult population in the world
64-bit
18 quintrillion
Grains of sand at a large beach
😱 😀
but please keep my last post in mind when reading this, it's a VERY MISLEADING generalisation! It just ain't that simple.
Besides, they also write: "Analysts say 64-bit computing will likely be in the mainstream by the year 2002 [note the date it was written, 1995 ], when operating systems and applications will be widely available."
oops.... 😉 😛
just found a cute comparison here, where they're talking about "A CPU's ability to address data depends on how many bits it can handle at one time":
8-bit
256
About 10 times larger than our alphabet
16-bit
65,536
Population of a midsize city
32-bit
4.3 billion
Total adult population in the world
64-bit
18 quintrillion
Grains of sand at a large beach
😱 😀
but please keep my last post in mind when reading this, it's a VERY MISLEADING generalisation! It just ain't that simple.
Besides, they also write: "Analysts say 64-bit computing will likely be in the mainstream by the year 2002 [note the date it was written, 1995 ], when operating systems and applications will be widely available."
oops.... 😉 😛
Re: Check your facts... and REALITY!
Quite true. However, there are enough apps that would see a speedup with 64 bits (primarily apps that use floating points now because 32-bit ints don't have the range and 64-bit ints on a 32-bit processor are too inefficient) that moving to 64-bit would be advantageous.
Correct me if I'm wrong, but the PPC 970 still uses 32-bit wide instructions, it just adds instructions for loading/manipulating/storing 64-bit ints alongside the 32-bit int load/manip/store instructions. As such, I don't believe you will "compile as a 32-bit binary"; you will compile with 64-bit integer support either "real" (970+ compatible only) or "simulated" (backwards-compatible to 32-bit PPC processors, using the same strategy for "long long" as gcc does currently). The "real" 64-bit binary will work better on the 970 than "simulated" 64-bits, but will not run on 32-bit processors. For the most part, your app will be compiled in both modes and will decide at runtime or install time which library (64-bit true or simulated) should be used based on the machine's processor (which is, not surprisingly, just what we do for AltiVec currently).
The code should not change size at all. I mean, the only place where I can see code size being larger with 64-bit int support is in your data segments, assuming that data you had previously defined as 32-bit wide is now defined as 64-bits wide. In general, though, the bulk of your code will not change at all if you are at least slightly careful in how you write your code.
And, yes, developers (at least the good ones 🙂 ) are completely capable of using 32-bit ints in a 64-bit world. Looking at my code, I've got quite a few 8-bit and 16-bit ints used, and 32-bit processors have been around since before I even started coding! There is a tendency in those who are new at it, to just use "int" when you only need to count to 10, but with experience and a few grueling performance optimizations using the right bit-width becomes second nature.
Yes, quite true. However, I think with the 970 optimizing for the Mac will suddenly get a lot harder. Up to this point, we've known the bottleneck 90% of the time is in the FSB. You just can't pull data out of memory fast enough to keep the processor busy. So cache management is the absolute highest priority, even if it leads to less efficient code at the CPU itself. With the 970, that gets turned on its head: the FSB is rip-roaring fast, and the bottleneck believe it or not is more akin to what you see on Intel processors: the CPU speed itself and the peripherals themselves, not the communication between them. While cache management is still important in this case, it takes a fairly quick back seat to keeping the CPU doing the "right" thing instead of wasting cycles.
The transition isn't going to be really pretty. If you don't optimize your code for the 970 architecture, it won't run a whole lot faster than it would on a similarly-clocked G4. However, if you optimize your code for the 970 it will be blazing fast on that new processor, and much slower on older G4s. You have to both know your customer and their machines, and have the ability to manage the transition along with them. In a performance-driven situation, 32-bit or 64-bit is really going to be the least of your worries!
Ahhh ... Yes, I remember saying that 16MB (on an Intel machine) should be more than enough for Word with just such a snicker on my face ... Now I wouldn't dream of loading up Word on a machine with less than 256MB ...
But, yes, hitting that 64-bit addressable memory boundary should take us a good long time. It's nice to have a wall move back away from us before we really hit it too hard!
Which sorta touches on the fact that Apple has a lot of work ahead of them in making OSX 64-bit friendly. Carbon's API, in general, will either remain 32-bit only or will require a massive amount of re-writing and dual-API support (all those structs with pointers now need 64-bit pointers!)
Thanks, I'm glad someone else sees it the same way!
Quite true. However, there are enough apps that would see a speedup with 64 bits (primarily apps that use floating points now because 32-bit ints don't have the range and 64-bit ints on a 32-bit processor are too inefficient) that moving to 64-bit would be advantageous.
Correct me if I'm wrong, but the PPC 970 still uses 32-bit wide instructions, it just adds instructions for loading/manipulating/storing 64-bit ints alongside the 32-bit int load/manip/store instructions. As such, I don't believe you will "compile as a 32-bit binary"; you will compile with 64-bit integer support either "real" (970+ compatible only) or "simulated" (backwards-compatible to 32-bit PPC processors, using the same strategy for "long long" as gcc does currently). The "real" 64-bit binary will work better on the 970 than "simulated" 64-bits, but will not run on 32-bit processors. For the most part, your app will be compiled in both modes and will decide at runtime or install time which library (64-bit true or simulated) should be used based on the machine's processor (which is, not surprisingly, just what we do for AltiVec currently).
The code should not change size at all. I mean, the only place where I can see code size being larger with 64-bit int support is in your data segments, assuming that data you had previously defined as 32-bit wide is now defined as 64-bits wide. In general, though, the bulk of your code will not change at all if you are at least slightly careful in how you write your code.
And, yes, developers (at least the good ones 🙂 ) are completely capable of using 32-bit ints in a 64-bit world. Looking at my code, I've got quite a few 8-bit and 16-bit ints used, and 32-bit processors have been around since before I even started coding! There is a tendency in those who are new at it, to just use "int" when you only need to count to 10, but with experience and a few grueling performance optimizations using the right bit-width becomes second nature.
Yes, quite true. However, I think with the 970 optimizing for the Mac will suddenly get a lot harder. Up to this point, we've known the bottleneck 90% of the time is in the FSB. You just can't pull data out of memory fast enough to keep the processor busy. So cache management is the absolute highest priority, even if it leads to less efficient code at the CPU itself. With the 970, that gets turned on its head: the FSB is rip-roaring fast, and the bottleneck believe it or not is more akin to what you see on Intel processors: the CPU speed itself and the peripherals themselves, not the communication between them. While cache management is still important in this case, it takes a fairly quick back seat to keeping the CPU doing the "right" thing instead of wasting cycles.
The transition isn't going to be really pretty. If you don't optimize your code for the 970 architecture, it won't run a whole lot faster than it would on a similarly-clocked G4. However, if you optimize your code for the 970 it will be blazing fast on that new processor, and much slower on older G4s. You have to both know your customer and their machines, and have the ability to manage the transition along with them. In a performance-driven situation, 32-bit or 64-bit is really going to be the least of your worries!
Ahhh ... Yes, I remember saying that 16MB (on an Intel machine) should be more than enough for Word with just such a snicker on my face ... Now I wouldn't dream of loading up Word on a machine with less than 256MB ...
But, yes, hitting that 64-bit addressable memory boundary should take us a good long time. It's nice to have a wall move back away from us before we really hit it too hard!
Which sorta touches on the fact that Apple has a lot of work ahead of them in making OSX 64-bit friendly. Carbon's API, in general, will either remain 32-bit only or will require a massive amount of re-writing and dual-API support (all those structs with pointers now need 64-bit pointers!)
Thanks, I'm glad someone else sees it the same way!
Dudemeister rant
Well about half of what was said is relevant. Remember that IBM is making these chips initially as a DESKTOP CPU for themselves.
If they wanted to make another Server CPU well.. they already have several. Obviously this chip is meant to fall into a reasonable price range for desktop systems. That includes all the I/O and Interconnect stuff as well...
So it is completely unimaginable that the 970 would not make it into Apple desktop systems within a reasonable amount of time, even if they decide to put them in XServes immediately for manufacturing testing purposes or other cost/benefit reasons. AND it will be at a decent pricepoint.
There are plenty of numbers posted so I won't rehash on performance.
I don't think anyone is really going to have an idea of the impact 64bit will have for the OS until it is seen in action, so speculation and pseudo-technical discussions on what it is or isn't good for are really unfounded at the moment.
Think back to 8bit and 16bit systems and the limitations they had that weren't even considered limitations at the time because no one had even contemplated using all the extra room 32bit has given us....
can you imagine doing what Quartz does now in any less than 32 bit? For G4s it also has altivec support, hardware acceleration, etc. What new innovative uses for 64bit might arise? Apple is known for pushing the envelope... I'm sure they've put some thought into how to take advantage of the extra resources, computational space, etc. that 64bit will allow for.
Well about half of what was said is relevant. Remember that IBM is making these chips initially as a DESKTOP CPU for themselves.
If they wanted to make another Server CPU well.. they already have several. Obviously this chip is meant to fall into a reasonable price range for desktop systems. That includes all the I/O and Interconnect stuff as well...
So it is completely unimaginable that the 970 would not make it into Apple desktop systems within a reasonable amount of time, even if they decide to put them in XServes immediately for manufacturing testing purposes or other cost/benefit reasons. AND it will be at a decent pricepoint.
There are plenty of numbers posted so I won't rehash on performance.
I don't think anyone is really going to have an idea of the impact 64bit will have for the OS until it is seen in action, so speculation and pseudo-technical discussions on what it is or isn't good for are really unfounded at the moment.
Think back to 8bit and 16bit systems and the limitations they had that weren't even considered limitations at the time because no one had even contemplated using all the extra room 32bit has given us....
can you imagine doing what Quartz does now in any less than 32 bit? For G4s it also has altivec support, hardware acceleration, etc. What new innovative uses for 64bit might arise? Apple is known for pushing the envelope... I'm sure they've put some thought into how to take advantage of the extra resources, computational space, etc. that 64bit will allow for.
Re: Dudemeister rant
Agreed. And remember, Apple is not going to want to support both 32-bit and 64-bit versions of their OS(es) any longer than they have to, so moving their entire product line to the 970 would make a lot of sense in that regard. Also, by moving the entire product line to the 970 Apple would be buying large numbers of one kind of chip (and therefore paying less for them) as opposed to buying smaller quantities of two, or even three different kinds of chips.
Re: Check your facts... and REALITY!
I disagree. We've seen other info that points to the same cost as the G4 if not cheaper. The 970 is a 'desktop' version of the power4, not a workstation version. I'd look for the current pricing to stay about the same. Apple has alos been very aggressive with their pricing, I expect that not to change.
No keynote at MWNY means no 970 release at MWNY. It will either be before then or after then probably at a special event.
I disagree. We've seen other info that points to the same cost as the G4 if not cheaper. The 970 is a 'desktop' version of the power4, not a workstation version. I'd look for the current pricing to stay about the same. Apple has alos been very aggressive with their pricing, I expect that not to change.
No keynote at MWNY means no 970 release at MWNY. It will either be before then or after then probably at a special event.
The main reason that the PIII is still used in servers is that the P4 Xeon generates far too much heat to be used in compact enclosures.
The 970 gives P4 comparable performance dissipating PIII comparable amounts of heat - it is "the best of both worlds." This is why IBM is planning to produce PPC 970 blades running Linux.
Re: Re: 970
OS 9 booting powermacs are still available, though I understand your point. My point was that 970's would be forever foreclosed from booting into OS 9.
OS 9 booting powermacs are still available, though I understand your point. My point was that 970's would be forever foreclosed from booting into OS 9.
Umm... relevance? 8))
Let me answer that: NO. Guess what, IBM intends them for Blades (meaning SERVERS)... sorry!
Well, I don't think I have the time to go into your definition of "reasonable" or "decent", since it quite obviously is way beyond my definition. And since both are mere speculation, there's really no POINT, is there?
But that is going out on too much of a tangent at this point, so let's not elaborate.
I'll refrain from offering a "relevance" grade myself, if that's okay with you.
cheerz
Well, sorry, but judging by what you've written, you only understood about half! So let's not fall into a quasi-flame deal here, okay?
Hmmm, as far as their website is concerned... I don't see much PPC desktop action going on (nor their notebbooks or workstations for that matter), so I'm not exactly sure what you could mean... I don't know of any plans on the part of IBM of fitting desktops with PPCs, do you?
Let me answer that: NO. Guess what, IBM intends them for Blades (meaning SERVERS)... sorry!
That's probably why I was talking about servers for APPLE... aside from the above of course. 🙄
Only IBM doesn't make nor SELL I/O components such as those that are needed. There IS a reason why server boxes are a might more expensive than desktops, huh? And it ain't just 'cause of the CPU... have you checked what e.g. RAM clocked at 400+ MHz costs lately? ... and just to cut the wallowing suspense, again, IBM has ZERO plans of putting the 970 in any DESKTOPS, okay? It's intended merely for "low-end" Blades!
Hmmm, completely UNimaginable that it would NOT? Two negatives DO make a positive...!
Well, I don't think I have the time to go into your definition of "reasonable" or "decent", since it quite obviously is way beyond my definition. And since both are mere speculation, there's really no POINT, is there?
I think I'll stick with a "no comment" on that one...
Sorry, but I dont think you know what you're talking about.
Such as? Apple has never pushed any envelopes that hadn't been pushed already that I can recall. I'd say they've kept up pretty well so far, except for the last couple of years, nothing more.
Good. Unfortunately, i.e. just like with OS X, as long as they're doing it by themselves and don't have any decent "playmates" tagging along, that effort has ZERO value. Which fortunately isn't the case the further we get along, but I remind you of others such as Be or even NeXT for that matter.
But that is going out on too much of a tangent at this point, so let's not elaborate.
I'll refrain from offering a "relevance" grade myself, if that's okay with you.
cheerz
Ummm.... *?!*
Okay.
Tho (I know it may have been a lot to cover) try not to only QUOTE the posts next time, but also READ them, it'll save you both time and typing... 😉
Okay.
Tho (I know it may have been a lot to cover) try not to only QUOTE the posts next time, but also READ them, it'll save you both time and typing... 😉
Re: Re: wwhaaaaaa...t?
no.
It's not a sales slump. PowerMacs are getting blown away by P4 based machines. Would you rather have to wait 2 seconds or 20 seconds for Maya to render a scene?
Compression??!
With an FSB running at 900MHz and ApplePI I think that there should be enough data to feed the PPC970. I think I have a solid grasp on the "meaning" of 64bit bus/CPU.
no.
It's not a sales slump. PowerMacs are getting blown away by P4 based machines. Would you rather have to wait 2 seconds or 20 seconds for Maya to render a scene?
Compression??!
With an FSB running at 900MHz and ApplePI I think that there should be enough data to feed the PPC970. I think I have a solid grasp on the "meaning" of 64bit bus/CPU.
-hh
macrumors 68030
Re: some people will still buy when they can
You're absolutely right.
The only real "flexibility" that large organizations have is what to spend their money on, within their spending window: it was for this reason alone that I was able to get $98K for a new camera system for my project last fiscal year.
For the PowerMac desktop, the only thing at this time that would likely "save" .EDU-type orders for Apple within their purchasing window would be a policy that Apple has been very reluctant to offer: a product preannouncment that's more than ~2 months out. The bean counters can usually be convinced to look the other way if the product has >30 day delivery, just so long as the money is obligated on time.
But to do this, Apple has to announce the product, price and Order#'s. Even if the actual physical delivery is still 3-4 months out.
-hh
You're absolutely right.
The only real "flexibility" that large organizations have is what to spend their money on, within their spending window: it was for this reason alone that I was able to get $98K for a new camera system for my project last fiscal year.
For the PowerMac desktop, the only thing at this time that would likely "save" .EDU-type orders for Apple within their purchasing window would be a policy that Apple has been very reluctant to offer: a product preannouncment that's more than ~2 months out. The bean counters can usually be convinced to look the other way if the product has >30 day delivery, just so long as the money is obligated on time.
But to do this, Apple has to announce the product, price and Order#'s. Even if the actual physical delivery is still 3-4 months out.
-hh
Jon the Heretic
macrumors 6502
Prior to the introduction of MacOS X, I have to admit I believed that a Mac faster than 500Mhz or so was pretty much useless for most users, other than games or video production.
However, Aqua has changed all of the rules. As the most sluggish GUI ever devised by man or fruit, I think a 2.5Ghz PPC 970 with a full 900Mhz FSB would provide an enormous value to ordinary users attempting to complete grueling everyday tasks on MacOS X, such as resizing a window or using the Finder to do things you could do on System 6.x with a MacPlus with alacrity.
Bring on the power!!! Yeah!!! Eventually, MacOS X will get that *snap* back that we used to expect from the 8Hz 68000 days. 2.5 Ghz is a damn good start. Now if only Nvidia will triple the clockspeed of their GPU to match, Aqua will have a chance to shine----in a translucent, glittery kind of way...
--------------------
Ooh, Shine-y. Gotta get me some of that. (Who needs a user friendly GUI when you got pretty colors that are shiny? It is a matter of priorities).
However, Aqua has changed all of the rules. As the most sluggish GUI ever devised by man or fruit, I think a 2.5Ghz PPC 970 with a full 900Mhz FSB would provide an enormous value to ordinary users attempting to complete grueling everyday tasks on MacOS X, such as resizing a window or using the Finder to do things you could do on System 6.x with a MacPlus with alacrity.
Bring on the power!!! Yeah!!! Eventually, MacOS X will get that *snap* back that we used to expect from the 8Hz 68000 days. 2.5 Ghz is a damn good start. Now if only Nvidia will triple the clockspeed of their GPU to match, Aqua will have a chance to shine----in a translucent, glittery kind of way...
--------------------
Ooh, Shine-y. Gotta get me some of that. (Who needs a user friendly GUI when you got pretty colors that are shiny? It is a matter of priorities).
Dont Hurt Me
macrumors 603
Apple will not need dual machines, i repeat this was only to make up for motorola and where it was not going and that didnt even do it for most apps. A single 1.8 970 will be about equal in power to a p4 running at about 3 ghz. This puts apple right on par with wintel period. Now this thing is scaling up so this means apple is right back in it for a long time. Give up your duallies, a 1.6 970 will run all over dual 1.42 so quit the duallie crap. Duallies wont be needed. Apple had to market something and motorola wasnt giving it. Do you Duallies people understand? This is great news if true. I for one will be thrilled if apple announce a 1.6 1.8 and 2.0 970. these will blow the doors off anything they have done. And please remember that the 970 will be able to use Ram after Ram. No longer limited to a puny 1.5 or 2 gigs. Do you all comprehend this? the G4 is a has been. Get over it. Apple wont need duallies for a very long time and neither will you. Duallies where smoke and mirrors for 90% of applications. enough said bring on the SINGLE 970.