Jimmac: you can call me anything you want, just don't call me late for dinner....
Anyway, this thread is just pissing in the piss of truth in the hurricane of hopes, and in the long run, since my position (that the G5 is only a figment of Ryan Meader's imagination) is sane, and the existence of the G5 is based entirely on hope (And fear of Apple's future), people will continue to hope for things that can't happen (world peace, end to the common cold, the G5, etc) there's not much I can do to alter perception. Heck, even the Register, which started fanning the "G5" claims last year has backed off, realizing that it's Apollo time, and hasn't even uttered G5 in months. However, they have been reporting on Apollo pretty much continuously.
Face it: processor design and fab is pretty much an open secret. We always know what's coming next at Intel, AMD, IBM and Motorola, because they give presentations and schedules and such at things like the Microproc Forum and such, and don't have the same super Iron Curtain around their products that Apple does around their round pieces of plastic. IBM's latest PPC is the Sahara, and we've known about it for almost a year. Intel's latest Nothwood (someone correct me if I'm wrong on the codename) just came out, but we've been hearing about it for months. AMD's XP chip was an open secret for a long time, too. The last major Moto chip (the 7450 and 7440 series) was known about for at least 8 months before introduction.
So, now comes Apollo, which has been on the public's chart for at least 8 months, if not more.
There is no way an entire new generation of proc is being hidden entirely from us. Since there is no evidence that the G5 has passed anything beyond the design phase, you might want to believe it.
Apple may be an intensely secret company, but Motorola is not and cannot be.
People who believe that the G5 is coming soon also believe that we didn't land on the moon, and that Oswald didn't shoot JFK. They take slight bits of scant information that stands out of phase with the overwhelming majority of the information that proves the event, then construct an elaborate fantasy by which things could be true. Then they change 'could' to 'is' because it suits their personal persuasion, and convince themselves that the scraps are the truth, and the material is a hoax.
That's why the standard of evidence most typically used is "beyond a reasonable doubt" and not "beyond a shadow of a doubt."
So, I can affirmatively say that, beyond a reasonable doubt, the Apollos is on it's way and will be here soon. I can also affirm that, beyond a reasonable doubt, the G5 is a research project more than a product.
I'm afraid I have to agree with ol' suckful here. I have seen NOTHING about the 8500 design being completed yet. All the reliable stories I've read refer to 0.x versions of the chip. It's probably too late for MWNY already, but the Apollos should come sooner rather than later - Apple says so. Their current rebate offer on the towers + flatscreens expires January 31st, so a February 1st announcement is indicated.
I am using the model number of the PPC fabbed by Motorola.
7460=G4
8500=G5
anything else=null and void.
...
SdC</strong><hr></blockquote>
I'm not offering any insider information here (I can't), but anybody participating in such a high-stakes bet should make sure that you opened it up to IBM as well. Clearly, you would have to cover your butt somehow against IBM possibly making a G4 or a G3 variation, but enough people seem to think that IBM will have a G5 contender that it would be a reasonable addition.
-yields and OSX were the hold ups as of late november
and my souce knows very little if any more at all.
Out of respect for my source who happens to be family I will only say that he/she is extremely credible and learned what they did in a professional setting from someone in cupertino. It was less of leak and more of a slip, but there was no *wink-nudge- wink* jerking around involved.
It is the only inside info i will probably ever come across.
<strong>Well, there is, but it's the 8500, and it's a ways off (2003).
What's coming next is the Apollo 7460 G4. It will be in Apple computers within the next 45 days.
Bookmark this thread, and come back to it. Take it to the bank, bet the farm, Katie bar the door, the chickens have come home to roost.
I guarantee it.
SdC</strong><hr></blockquote>
Erm, why does everyone speculate about the G5's desktop release using the 8540 as a guide? You do realize that the 8540 is much more complicated than the 8500?
As for the 8500 being here by 2003, conservative guesses are just as speculative as optimistic ones. Do you have any insider info worth repeating?
If so, I've been known to make a wager now and then.
Just to define the proposed terms:
I am using the model number of the PPC fabbed by Motorola.
7460=G4
8500=G5
anything else=null and void.
I'm not about to take a flyer on Apple's marketing folks (they may choose to call a G4 generation chip a G5, even if it's not).
Let me know if you really wanna take this bet.
For the rest of you; show me one non-rumor site source that says the G5 is anywhere NEAR ready.
Because I'm using The Register, Motorola, the Microprocessor Forum, and other reputable sources to assert my claim that Apollo is coming.
SdC</strong><hr></blockquote>
I'll take that bet, and I agree with your adding the no repackaged G4 called a G5. But just to not be specific (you list the 8500) why don't we just say a 8xxx series chip.
The more and more I look at the 8540 the more convinced a G5 this year is guaranteed. It does have a SIMD unit, though they don't refer to it as Altivec and the SIMD unit has 222 new instrunctions (where did I hear rumors about AltiveII?), three 128b data busses.
The MPC 8XXX series is listed as "Integrated Host Processors" which is surprisingly similar to the 7XXX, 7XX, & 6XX which are listed as "Host Processors". The 82XX and 8XX are listed as "Integrated Communications Processors".
Here's why the G5 will be released (let's also say that and not just announced, so there is no confussion) this year.
1. The 8540 (yes I know it's most likely not able to function as a Mac processor) is a great starting point. It is "Book E" compliant which makes it Mac compliant. It has support for APU's (Auxillary Processor Units) which is what Altivec is. If the SP (Signal Processing) APU on the chip is not already Altivec or Altivec II, it can be swapped for it as stated above (Book E compliant). This could go on forever, just look at all the available info on this chip and the current G4's to see the compatibility and natural progressions. If the 8540 can see the light of day in 2002 a G5 most certainly will. I'd bet the cores are nearly identical.
2. Actually the e500 core will be the core of the G5, here is the exerpt from Motorola's 8540 fact sheet........
[quote]
e500 Core
Utilizing an SoC platform which balances MIPs, watts, packet performance and cost, Motorola has created a flexible platform architecture enabling multiple products from easily integrated IP. The e500 high performance core implements the enhanced PowerPC "Book E" instruction set architecture and provides unprecedented levels of hardware and software debug support. The e500 will serve as the core for a family of ASSPs for communications, automotive and consumer applications. <hr></blockquote>
Of course I could be wrong about this.....You ask "show me one non-rumor site source that says the G5 is anywhere NEAR ready.
Because I'm using The Register, Motorola, the Microprocessor Forum, and other reputable sources to assert my claim that Apollo is coming."
[quote] The company also held off adopting G5 PowerPCs made in a 0.13-micron process. Those chips will sport a longer pipeline and hit speeds up to 1.4 GHz, but are not expected to be available in volume from Motorola until midyear. <hr></blockquote>
*******************STOP********************
I just had a thought, the G5 was originally believed to be numbered 75XX as the 8XX & 8XXX (8 series) are considered more embedded chips. So it would be possible that a core such as the e500 could be the "G5" core and used in both 8XXX and 75XX series. This has just become way more complicated than I thought.
I will take the bet, but we can not go by either part numbers or just the fact they call it a G5. We would have to look at the processor and what differientiates it from the G4. Any real architectural differences, not to merely include additional pipelines or larger on-die cache, but true changes that we could look at and agree would make a chip truly a G5 or just a enhanced G4.
Well, to be brutally honest, my bluff has been called. I'm holding a pair of dueces in the 'balls' department.
The fact that I am a coward gambler shouldn't change that I agree that it might get stickier than we think, because of the e500 spec, and the fact that the 8540 has a space for the auxillary processing stuff (which I did not know, primarily because I din't quite understand the evasive Motorola language).
In theory, there might be a "G5-class" processor coming out before the end of the year, although even with the 8540 just going into sampling in @H 2002, that's no sure thing, IMO. And even if the Apple-used G5 class proc was based on the 8540, it would have to go into sampling, and ramp up to high speeds very quickly to make it into Apple products; it is currently expected to cover the gamut from 600 MHz to 1 GHz. Given that even the current G4 could probably make 1 GHz by the end of the year just throu maturation of the fab process, it seems unlikely that a chip that is scheduled to debut lower than current chip is designed for Apple products.
Now, we move into the realm of speculation, but let's assume that the G5-class processor that Apple would be interested in is the 8540 PLUS the auxillary processing units (AltiVec) PLUS SOI PLUS a die shrink. That's a lot to expect by the end of the year, no?
Basically, though, you've busted my hyperbole. The intent of this thread was to counter all of the threads that said the G5 was ready today, and coming off the fabs and that Apple is holding them back for its own nefarious plans. It is possible (though unlikely, IMO) that a G5 class chip that Apple would use might be released by the end of the year (and I'm talking winter equinox possible), and the possibility is such that I'm not really willing to put $1700 on it.
But I think you would agree to this, which is really the point of the thread, hyperbole aside:
-That Apple will upgrade its Powermac line
-That Motorola is making the 7460, and is expected to be able to ship these in the next few months
-And that Apple will use the Motorola 7460 in the Powermac line before the 8xxx processor, simply because it is available sooner, and faster than the current 7451.
That having been said, I am still skeptical about an 8xxx processor in an Apple product before the end of the year, and am willing to put something of lesser value on the line, by about an order of magnitude or so.
Got any suggestions? I think the iPod (or other iThing that Apple may choose to create in the same price range) makes a wonderful 2002 Xmas gift, no?
My proposed terms would be that:
Apple would not have a product (that is to say, a Powermac, Powerbook, iMac, or iBook; they might invent a as-yet-unknown-new product for which this chip is suited for) shipping by the end of the year whose processor is based on 8xxx (as desmonstrated by the 8540) e500 spec.
How about that?
As for the eeTimes, I think they are suffering under the same confusion that has spawned this confusion in the first place: just what is a G4 and what is a G5? I think Motorola is being cagey on this, because (aside from the 8540, which must already have a target market which wants them really bad; I suspect routers, of which I really know nothing about) they haven't really even decided what they are doing about the generational divisions of their procs.
There is a lot of turmoil over at Moto around the Semiconductor biz, and they may be putting off generational road maps and plans because they honestly don't know what the future portends. (All the more reason for Apple to seek it's own destiny on the cpu, but that's neither here nor now.)
When Mot's 130nm process is ready to go online then the G5 will be a go. That's all the e500 core is waiting for at this point. It's too big to produce on 180nm, especially with all the desktop parts it needs. With a desktop part Mot doesn't have to worry about power consumption and heat generated. Thats what you worry about when it's an embedded chip. If the 8540 (the first embedded G5) will run 600-1000MHz at 6.5 watts (for the 600MHz I assume) then at 15 watts (not too bad for a desktop processor) it should run at at least 1200MHz. 20 watts around 1600MHz. And remember the desktop G5 will not have all the baggage the 8540 has like (2) Gb ethernet controllers, a 64bit PCI-X controller, serial port controllers, etc.
I think the 8500 will be the first G5 in the family. It will be a desktop processor for Apple almost exclusively because Apple will be footing most of the bill for the development of that processor. Since it's (the 8500) a Mot/Apple joint development and the 8540 isn't, Mot can pre-announce the embedded version but they have to keep mum on the 8500. At least that's my percieved explanation.
All th things you read on Motorola's website is things we can know; like new technologies they will use. In their Roadmap stands that the G5 will start at 800 MHz,maybe this is just to mislead us. Maybe at the time of writing this there is a G6 been tested. Somethings that can have a big impact on consumers are kept secret. Every rumor site is actually just guessing,or there is a major leak or a good source. Most of the time we will just have to wait until they show there new goodies.Good example is the iMac, everybody thaugt a sahara G3. Big bummer was it.
Personaly I guess the G? will certainly break the GHz barrier and even have some new features (USB2, gigawire,?)
You see, just wait and see. We 'll find out sooner or later.
Er, I hate to break this to you but,........well despite everything you read on the web and elsewhere...........these companies you have been talking about ( this is going to be tough so brace yourself ) DON'T REALLY TELL YOU EVERYTHING ABOUT WHAT THEY ARE DOING!
Every company does this. Companies like Apple and Motorola have vast resources to keep these things secret if they want.
I'm sorry but, I'm willing to bet there are lots of things Intel and Microsoft are doing that we know nothing about until they choose to let us know. Also companies like Microsoft can be a little more open because they aren't struggling ( besides they are the ones who copy or buy the labour of others ).
When you hear something on a rumour site that turns out to be true it's usually because somebody slipped up. That's happening less and less since the worker bee incident. When SJ is talking about the " Digital Hub " in SF Bill Gates is saying the same thing elsewhere only expanded on. Apple wants to avoid this sort of thing so they keep quiet.
It's not so much that SJ doesn't like his thunder taken away, he just doesn't like someone else doing the same thing before he gets a chance to get the product ( or idea ) out the door. If you don't believe me just look around ( remember all the imac ripp offs ).
So none of us are really privy to everything Apple or Motorla does ( or plans ).
1) Apple has taken over much (if not all) of the design responsiblities for the G5 project (Motorola has other things to worry about).
2) IBM cannot fab chips beyond the G3, not unless Steve is willing to abandon AltiVec (not bloody likely) or Moto licenses AltiVec (also not likely). While the Apple-IBM-Motorola former-alliance share the rights for the PowerPC (though IBM holds the tradmark on the name), Moto owns AltiVec all by itself. If this weren't the case, you can be sure that IBM would have been fabbing G4s long ago (and we'd have broken the GHz barrier a long time ago. Sorry Moto, but it's true...
I don't expect to see a G5 (whatever its designation may be) before July, though I'm sure will see 1GHz+ Pro models with 7460s at or before Seybold next month (hence the reason the Crystal Clear promotion was extended until January 31). I expect that Apple wants a healthy stockpile of high-speed Apollo chips before they announce and Motos recent (er, ongoing) yield problems probably set things back a little bit. If they look anything like the descriptions indicate, they're going to be as beautiful as they are fast.
As much as I'd love to see a 2+ GHz G5 show up out of the blue, it's not realistic.
I must admit this has been one of the best threads I've seen the the new AI. Similar to some of the old 1999 threads of classic AI.
Well thought out responses, a willingness to admit the other person(s) has made a valid point, and none (almost) of the childish whining, wild (and inconceivable) ranting, and utter "blah" that almost makes me hate this place.
Sometimes I think its almost useless to even try finding a good thread here. Seems like most threads are made up of "I want..., I want..., I want..." without any sense of practicality or reasonability. AAPL does not walk on water. [Your] ideas are not necessarily the best ideas.
Thanks for reinstilling confidence of the intellience of (some) of the people that make up the AI community.
That EE Times.com article is the first reference from what I think is not your typical rumor mill site mentioning anything about release dates for the G5.
If true, now my worry is that the Apollo may not make an appearance in the PowerMac at all and the PowerMacs will linger @ current speeds until the G5 is introduced.
"..the first RapidIO interconnect on a microprocessor.."
"..first processor to utilize the powerful e500 core, .."
"...the MPC8540 addresses the next generation needs of our customers, and provides Motorola with an ideal foundation for continued system-on-chip innovations."
As for the MPC8540 being more complicated than a desktop version, I can not believe that. This chip is "Designed for high-speed networking applications, the MPC8540 offers exceptional bandwidth built upon an optimized system-on-chip platform" A desktop cpu is much more generalized than this specialized chip and the desktop version I would think will be substatially larger with millions more transistors. Maybe some one with more knowledge of microprocessors can explain?
I'm not going to claim to be an expert on processors (or hardware, for that matter), but let me offer a few possibility:
Suppose the Apollo G4s have a bus that's tweaked to be compatible with the G5's, so Apple can put it on a daughtercard and attach it to a G5-compatible mobo (with DDR RAM, etc.). This means that they can release a new PowerMac soon, and then update the line at Seybold SF (say) with minimal reengineering. moki has stated that an Apollo-based DDR PowerMac does exist, although it's not quite finished yet.
Alternately (and less attractively), Apple could offer a speedbump to the existing line based on the Apollo chips now and introduce a whole new machine later this year.
(I'm assuming that the G5 will appear sometime midyear - which, to be honest, I'm not sure I believe either! If not, the Apollo will get another boost from the transition to a 0.13 micron process in the mean time).
My impression of the G4 is that it isn't very different from the G3. I also understand that the Apollo G4 is a bigger step up from the 'regular' G4 than the G4 was from the G3. From what I understand about the G5 though, it's significantly different than even the Apollo G4, albeit still compatible.
So, when the original G4 was introduced, it was at the same speeds as the G3, only with Altivec and an error that kept it from scaling up, hence the stuck at the same MHZ forever problem.
The G5 design doesn't tie in very closely with the G4. To me, this means its speed doesn't tie in very closely with that of the G4 either. So, following moore's law, had the G4 been introduced without the error, it would have scaled probably as high as the Athlons are now.
When introduced, the G5 could easily make a jump that large. It's not as if Motorola doesn't know how to make a chip go 1.6 GHZ, I think it's that they don't know how to make the G4 go that fast. There's a difference.
So, I don't think fast G5s are a pipedream. I think the problems the G4 has always had point to the fact that Apple and Motorola have been working on the problem for the next generation chip rather than fixing the G4 just in time for it's demise.
The change from 750 (G3) -> 7400 (G4) was a bit more than just adding AltiVec, but not much. The change from 7400 -> 7450 (G4+) was more extensive -- pipelines were lengthened and an on-chip L2 cache was added. Other improvements to support these changes were added. The 7450 -> 7460 (Apollo) change was supposed to be just an SOI & process shrink with a few internal optimizations, it remains to be seen if there are any bus improvements to support DDR or core changes. The last I heard though was that 7460 didn't include the process shrink, just the SOI and the process shrink would come later. That will limit how much of a clock rate jump we see in the near-term (barring the surprise arrival of the G5).
The G5 will be an entirely new core, much like the Pentium III -> Pentium IV transition. I agree with you that this will likely result in a large dicontinuity in performance from one machine to the next. The faster the Apollo turns out to be, the less the discontinuity will be.
Apple has been silent on the high-end performance front for some time now. They know they are falling quite a way behind, I just hope that they can do something about it soon.
I am willing to go smaller (I may have an iPod by that time but a 120GB firewire HD is about the same cost).
The G5 as pertains strictly to Apple has been in development for a minimum of 2 years. That can clearly be verified in Motorola press releases and presentations at trade shows. The problem is that shortly after they announced the G5 they went into silent running mode.
Truth be told is that Apple is footing alot of the R&D on the G5 and has it wrapped in SJ's RDF. This is probably the biggest simple reason you never here mention of the G5 except from "inside sources".
So Motorola is left with only one option to detail it's chips to the technology world. They have to talk about chips of the same class but that are for other markets (8540). It's a way to show the industry what your doing (blowing your own horn), without giving away Apple's precious secrets.
I am not guaranteeing that the e500 core is going to be the same core as the G5, but it sure looks like it could be. This core can have all kinds of APU's bolted on (which is part of the Book E standards) and they include Altivec. They could also drop (as someone suggested) things not necessary in the e500 core (serial int support, dual ethernet, can't remember what they suggested). That would only be necessary if these items took up any amount of die size (which I doubt most do).
As for the speeds listed, they are speeds specific to the 8540 product, not to any other product that may use the same core or a variant of it. Once the core (e500 in this case) is near completion, you begin to make chips for different sectors that utilize the core. Different performance levels, different applications. It is a modular design so you don't need to have the core 100% complete. In the same aspect you can be working on a G5 for Apple at the same time as it is a family class of processors that will all share similar aspects.
Now does all this "reasoning" (ha) imply that a G5 will be forthcoming this year? Heck no. So why do I think a G5 this year (and I'm thinking by July at the abosulte latest, but it may be earlier).
1. Wasn't Motorola's original time table for late '01 or early '02.
2. Motorola is currently shipping products on their HiPerMOS7 (Seventh Generation High Performance Metal Oxide Semiconductor) .13 micron lithography and SOI (Silicon on Insulator) technology along with copper interconnects.
3. Motorola says the 8540 will sample in 2nd half of 2002. Since the 8540 was developed at Somerset (where Apples chips are developed) and Apple foots part of the bill there, I would expect to see an Apple G5 class processor ship before an embedded Motorola part.
4. If I lose that would cut into my Xmas 2002 in Vegas fund (what can I say, my family likes to party). I need those coins for the slots and those greenbacks for the tables.
Also that with the addition of G4 to the iMac line I believe they need to differentiate the PowerMacs with a better processor. How long was the PowerMac and iMac all at G3?????
<strong>If this weren't the case, you can be sure that IBM would have been fabbing G4s long ago (and we'd have broken the GHz barrier a long time ago. Sorry Moto, but it's true...</strong><hr></blockquote>
Yes, IBM really boosted their G3 to GHz+ speeds - NOT!
[quote]If this weren't the case, you can be sure that IBM would have been fabbing G4s long ago (and we'd have broken the GHz barrier a long time ago. Sorry Moto, but it's true...<hr></blockquote>
ibm WAS fabbing the g4. the didn't change the design though, just fabbing...heard that from an apple emp at wwdc.
Comments
Anyway, this thread is just pissing in the piss of truth in the hurricane of hopes, and in the long run, since my position (that the G5 is only a figment of Ryan Meader's imagination) is sane, and the existence of the G5 is based entirely on hope (And fear of Apple's future), people will continue to hope for things that can't happen (world peace, end to the common cold, the G5, etc) there's not much I can do to alter perception. Heck, even the Register, which started fanning the "G5" claims last year has backed off, realizing that it's Apollo time, and hasn't even uttered G5 in months. However, they have been reporting on Apollo pretty much continuously.
Face it: processor design and fab is pretty much an open secret. We always know what's coming next at Intel, AMD, IBM and Motorola, because they give presentations and schedules and such at things like the Microproc Forum and such, and don't have the same super Iron Curtain around their products that Apple does around their round pieces of plastic. IBM's latest PPC is the Sahara, and we've known about it for almost a year. Intel's latest Nothwood (someone correct me if I'm wrong on the codename) just came out, but we've been hearing about it for months. AMD's XP chip was an open secret for a long time, too. The last major Moto chip (the 7450 and 7440 series) was known about for at least 8 months before introduction.
So, now comes Apollo, which has been on the public's chart for at least 8 months, if not more.
There is no way an entire new generation of proc is being hidden entirely from us. Since there is no evidence that the G5 has passed anything beyond the design phase, you might want to believe it.
Apple may be an intensely secret company, but Motorola is not and cannot be.
People who believe that the G5 is coming soon also believe that we didn't land on the moon, and that Oswald didn't shoot JFK. They take slight bits of scant information that stands out of phase with the overwhelming majority of the information that proves the event, then construct an elaborate fantasy by which things could be true. Then they change 'could' to 'is' because it suits their personal persuasion, and convince themselves that the scraps are the truth, and the material is a hoax.
That's why the standard of evidence most typically used is "beyond a reasonable doubt" and not "beyond a shadow of a doubt."
So, I can affirmatively say that, beyond a reasonable doubt, the Apollos is on it's way and will be here soon. I can also affirm that, beyond a reasonable doubt, the G5 is a research project more than a product.
Don't confuse the ghost for the body.
SdC
<strong>
You really wanna take that bet?
...
I am using the model number of the PPC fabbed by Motorola.
7460=G4
8500=G5
anything else=null and void.
...
SdC</strong><hr></blockquote>
I'm not offering any insider information here (I can't), but anybody participating in such a high-stakes bet should make sure that you opened it up to IBM as well. Clearly, you would have to cover your butt somehow against IBM possibly making a G4 or a G3 variation, but enough people seem to think that IBM will have a G5 contender that it would be a reasonable addition.
I know VERY little about the g5 other than
-it will be 64 bit
-yields and OSX were the hold ups as of late november
and my souce knows very little if any more at all.
Out of respect for my source who happens to be family I will only say that he/she is extremely credible and learned what they did in a professional setting from someone in cupertino. It was less of leak and more of a slip, but there was no *wink-nudge- wink* jerking around involved.
It is the only inside info i will probably ever come across.
We'll have to just wait and see.
[ 01-09-2002: Message edited by: jj ]</p>
<strong>Well, there is, but it's the 8500, and it's a ways off (2003).
What's coming next is the Apollo 7460 G4. It will be in Apple computers within the next 45 days.
Bookmark this thread, and come back to it. Take it to the bank, bet the farm, Katie bar the door, the chickens have come home to roost.
I guarantee it.
SdC</strong><hr></blockquote>
Erm, why does everyone speculate about the G5's desktop release using the 8540 as a guide? You do realize that the 8540 is much more complicated than the 8500?
As for the 8500 being here by 2003, conservative guesses are just as speculative as optimistic ones. Do you have any insider info worth repeating?
<strong>
You really wanna take that bet?
If so, I've been known to make a wager now and then.
Just to define the proposed terms:
I am using the model number of the PPC fabbed by Motorola.
7460=G4
8500=G5
anything else=null and void.
I'm not about to take a flyer on Apple's marketing folks (they may choose to call a G4 generation chip a G5, even if it's not).
Let me know if you really wanna take this bet.
For the rest of you; show me one non-rumor site source that says the G5 is anywhere NEAR ready.
Because I'm using The Register, Motorola, the Microprocessor Forum, and other reputable sources to assert my claim that Apollo is coming.
SdC</strong><hr></blockquote>
I'll take that bet, and I agree with your adding the no repackaged G4 called a G5. But just to not be specific (you list the 8500) why don't we just say a 8xxx series chip.
The more and more I look at the 8540 the more convinced a G5 this year is guaranteed. It does have a SIMD unit, though they don't refer to it as Altivec and the SIMD unit has 222 new instrunctions (where did I hear rumors about AltiveII?), three 128b data busses.
The MPC 8XXX series is listed as "Integrated Host Processors" which is surprisingly similar to the 7XXX, 7XX, & 6XX which are listed as "Host Processors". The 82XX and 8XX are listed as "Integrated Communications Processors".
Here's why the G5 will be released (let's also say that and not just announced, so there is no confussion) this year.
1. The 8540 (yes I know it's most likely not able to function as a Mac processor) is a great starting point. It is "Book E" compliant which makes it Mac compliant. It has support for APU's (Auxillary Processor Units) which is what Altivec is. If the SP (Signal Processing) APU on the chip is not already Altivec or Altivec II, it can be swapped for it as stated above (Book E compliant). This could go on forever, just look at all the available info on this chip and the current G4's to see the compatibility and natural progressions. If the 8540 can see the light of day in 2002 a G5 most certainly will. I'd bet the cores are nearly identical.
2. Actually the e500 core will be the core of the G5, here is the exerpt from Motorola's 8540 fact sheet........
[quote]
e500 Core
Utilizing an SoC platform which balances MIPs, watts, packet performance and cost, Motorola has created a flexible platform architecture enabling multiple products from easily integrated IP. The e500 high performance core implements the enhanced PowerPC "Book E" instruction set architecture and provides unprecedented levels of hardware and software debug support. The e500 will serve as the core for a family of ASSPs for communications, automotive and consumer applications. <hr></blockquote>
Of course I could be wrong about this.....You ask "show me one non-rumor site source that says the G5 is anywhere NEAR ready.
Because I'm using The Register, Motorola, the Microprocessor Forum, and other reputable sources to assert my claim that Apollo is coming."
Well here you go:
<a href="http://www.eetimes.com/story/OEG20020108S0030"; target="_blank">http://www.eetimes.com/story/OEG20020108S0030</a>;
[quote] The company also held off adopting G5 PowerPCs made in a 0.13-micron process. Those chips will sport a longer pipeline and hit speeds up to 1.4 GHz, but are not expected to be available in volume from Motorola until midyear. <hr></blockquote>
*******************STOP********************
I just had a thought, the G5 was originally believed to be numbered 75XX as the 8XX & 8XXX (8 series) are considered more embedded chips. So it would be possible that a core such as the e500 could be the "G5" core and used in both 8XXX and 75XX series. This has just become way more complicated than I thought.
I will take the bet, but we can not go by either part numbers or just the fact they call it a G5. We would have to look at the processor and what differientiates it from the G4. Any real architectural differences, not to merely include additional pipelines or larger on-die cache, but true changes that we could look at and agree would make a chip truly a G5 or just a enhanced G4.
What do you say?
The fact that I am a coward gambler shouldn't change that I agree that it might get stickier than we think, because of the e500 spec, and the fact that the 8540 has a space for the auxillary processing stuff (which I did not know, primarily because I din't quite understand the evasive Motorola language).
In theory, there might be a "G5-class" processor coming out before the end of the year, although even with the 8540 just going into sampling in @H 2002, that's no sure thing, IMO. And even if the Apple-used G5 class proc was based on the 8540, it would have to go into sampling, and ramp up to high speeds very quickly to make it into Apple products; it is currently expected to cover the gamut from 600 MHz to 1 GHz. Given that even the current G4 could probably make 1 GHz by the end of the year just throu maturation of the fab process, it seems unlikely that a chip that is scheduled to debut lower than current chip is designed for Apple products.
Now, we move into the realm of speculation, but let's assume that the G5-class processor that Apple would be interested in is the 8540 PLUS the auxillary processing units (AltiVec) PLUS SOI PLUS a die shrink. That's a lot to expect by the end of the year, no?
Basically, though, you've busted my hyperbole. The intent of this thread was to counter all of the threads that said the G5 was ready today, and coming off the fabs and that Apple is holding them back for its own nefarious plans. It is possible (though unlikely, IMO) that a G5 class chip that Apple would use might be released by the end of the year (and I'm talking winter equinox possible), and the possibility is such that I'm not really willing to put $1700 on it.
But I think you would agree to this, which is really the point of the thread, hyperbole aside:
-That Apple will upgrade its Powermac line
-That Motorola is making the 7460, and is expected to be able to ship these in the next few months
-And that Apple will use the Motorola 7460 in the Powermac line before the 8xxx processor, simply because it is available sooner, and faster than the current 7451.
That having been said, I am still skeptical about an 8xxx processor in an Apple product before the end of the year, and am willing to put something of lesser value on the line, by about an order of magnitude or so.
Got any suggestions?
My proposed terms would be that:
Apple would not have a product (that is to say, a Powermac, Powerbook, iMac, or iBook; they might invent a as-yet-unknown-new product for which this chip is suited for) shipping by the end of the year whose processor is based on 8xxx (as desmonstrated by the 8540) e500 spec.
How about that?
As for the eeTimes, I think they are suffering under the same confusion that has spawned this confusion in the first place: just what is a G4 and what is a G5? I think Motorola is being cagey on this, because (aside from the 8540, which must already have a target market which wants them really bad; I suspect routers, of which I really know nothing about) they haven't really even decided what they are doing about the generational divisions of their procs.
There is a lot of turmoil over at Moto around the Semiconductor biz, and they may be putting off generational road maps and plans because they honestly don't know what the future portends. (All the more reason for Apple to seek it's own destiny on the cpu, but that's neither here nor now.)
SdC
I think the 8500 will be the first G5 in the family. It will be a desktop processor for Apple almost exclusively because Apple will be footing most of the bill for the development of that processor. Since it's (the 8500) a Mot/Apple joint development and the 8540 isn't, Mot can pre-announce the embedded version but they have to keep mum on the 8500. At least that's my percieved explanation.
Personaly I guess the G? will certainly break the GHz barrier and even have some new features (USB2, gigawire,?)
You see, just wait and see. We 'll find out sooner or later.
Er, I hate to break this to you but,........well despite everything you read on the web and elsewhere...........these companies you have been talking about ( this is going to be tough so brace yourself ) DON'T REALLY TELL YOU EVERYTHING ABOUT WHAT THEY ARE DOING!
Every company does this. Companies like Apple and Motorola have vast resources to keep these things secret if they want.
I'm sorry but, I'm willing to bet there are lots of things Intel and Microsoft are doing that we know nothing about until they choose to let us know. Also companies like Microsoft can be a little more open because they aren't struggling ( besides they are the ones who copy or buy the labour of others ).
When you hear something on a rumour site that turns out to be true it's usually because somebody slipped up. That's happening less and less since the worker bee incident. When SJ is talking about the " Digital Hub " in SF Bill Gates is saying the same thing elsewhere only expanded on. Apple wants to avoid this sort of thing so they keep quiet.
It's not so much that SJ doesn't like his thunder taken away, he just doesn't like someone else doing the same thing before he gets a chance to get the product ( or idea ) out the door. If you don't believe me just look around ( remember all the imac ripp offs ).
So none of us are really privy to everything Apple or Motorla does ( or plans ).
Thus endth the lesson.
[ 01-09-2002: Message edited by: jimmac ]</p>
1) Apple has taken over much (if not all) of the design responsiblities for the G5 project (Motorola has other things to worry about).
2) IBM cannot fab chips beyond the G3, not unless Steve is willing to abandon AltiVec (not bloody likely) or Moto licenses AltiVec (also not likely). While the Apple-IBM-Motorola former-alliance share the rights for the PowerPC (though IBM holds the tradmark on the name), Moto owns AltiVec all by itself. If this weren't the case, you can be sure that IBM would have been fabbing G4s long ago (and we'd have broken the GHz barrier a long time ago. Sorry Moto, but it's true...
I don't expect to see a G5 (whatever its designation may be) before July, though I'm sure will see 1GHz+ Pro models with 7460s at or before Seybold next month (hence the reason the Crystal Clear promotion was extended until January 31). I expect that Apple wants a healthy stockpile of high-speed Apollo chips before they announce and Motos recent (er, ongoing) yield problems probably set things back a little bit. If they look anything like the descriptions indicate, they're going to be as beautiful as they are fast.
As much as I'd love to see a 2+ GHz G5 show up out of the blue, it's not realistic.
Well thought out responses, a willingness to admit the other person(s) has made a valid point, and none (almost) of the childish whining, wild (and inconceivable) ranting, and utter "blah" that almost makes me hate this place.
Sometimes I think its almost useless to even try finding a good thread here. Seems like most threads are made up of "I want..., I want..., I want..." without any sense of practicality or reasonability. AAPL does not walk on water. [Your] ideas are not necessarily the best ideas.
Thanks for reinstilling confidence of the intellience of (some) of the people that make up the AI community.
mmm chicken!
That EE Times.com article is the first reference from what I think is not your typical rumor mill site mentioning anything about release dates for the G5.
If true, now my worry is that the Apollo may not make an appearance in the PowerMac at all and the PowerMacs will linger @ current speeds until the G5 is introduced.
I have $1700 NOW and badly need to upgrade.
<a href="http://www.motorola.com/mediacenter/news/detail/0,1958,568_322_23,00.html"; target="_blank">http://www.motorola.com/mediacenter/news/detail/0,1958,568_32 2_23,00.html</a>
"..First e500 Integrated Host Processor.."
".. the MPC8540 is Motorola's first processor .."
"..the first RapidIO interconnect on a microprocessor.."
"..first processor to utilize the powerful e500 core, .."
"...the MPC8540 addresses the next generation needs of our customers, and provides Motorola with an ideal foundation for continued system-on-chip innovations."
As for the MPC8540 being more complicated than a desktop version, I can not believe that. This chip is "Designed for high-speed networking applications, the MPC8540 offers exceptional bandwidth built upon an optimized system-on-chip platform" A desktop cpu is much more generalized than this specialized chip and the desktop version I would think will be substatially larger with millions more transistors. Maybe some one with more knowledge of microprocessors can explain?
[ 01-09-2002: Message edited by: rickag ]</p>
Suppose the Apollo G4s have a bus that's tweaked to be compatible with the G5's, so Apple can put it on a daughtercard and attach it to a G5-compatible mobo (with DDR RAM, etc.). This means that they can release a new PowerMac soon, and then update the line at Seybold SF (say) with minimal reengineering. moki has stated that an Apollo-based DDR PowerMac does exist, although it's not quite finished yet.
Alternately (and less attractively), Apple could offer a speedbump to the existing line based on the Apollo chips now and introduce a whole new machine later this year.
(I'm assuming that the G5 will appear sometime midyear - which, to be honest, I'm not sure I believe either! If not, the Apollo will get another boost from the transition to a 0.13 micron process in the mean time).
My impression of the G4 is that it isn't very different from the G3. I also understand that the Apollo G4 is a bigger step up from the 'regular' G4 than the G4 was from the G3. From what I understand about the G5 though, it's significantly different than even the Apollo G4, albeit still compatible.
So, when the original G4 was introduced, it was at the same speeds as the G3, only with Altivec and an error that kept it from scaling up, hence the stuck at the same MHZ forever problem.
The G5 design doesn't tie in very closely with the G4. To me, this means its speed doesn't tie in very closely with that of the G4 either. So, following moore's law, had the G4 been introduced without the error, it would have scaled probably as high as the Athlons are now.
When introduced, the G5 could easily make a jump that large. It's not as if Motorola doesn't know how to make a chip go 1.6 GHZ, I think it's that they don't know how to make the G4 go that fast. There's a difference.
So, I don't think fast G5s are a pipedream. I think the problems the G4 has always had point to the fact that Apple and Motorola have been working on the problem for the next generation chip rather than fixing the G4 just in time for it's demise.
The G5 will be an entirely new core, much like the Pentium III -> Pentium IV transition. I agree with you that this will likely result in a large dicontinuity in performance from one machine to the next. The faster the Apollo turns out to be, the less the discontinuity will be.
Apple has been silent on the high-end performance front for some time now. They know they are falling quite a way behind, I just hope that they can do something about it soon.
I am willing to go smaller (I may have an iPod by that time but a 120GB firewire HD is about the same cost).
The G5 as pertains strictly to Apple has been in development for a minimum of 2 years. That can clearly be verified in Motorola press releases and presentations at trade shows. The problem is that shortly after they announced the G5 they went into silent running mode.
Truth be told is that Apple is footing alot of the R&D on the G5 and has it wrapped in SJ's RDF. This is probably the biggest simple reason you never here mention of the G5 except from "inside sources".
So Motorola is left with only one option to detail it's chips to the technology world. They have to talk about chips of the same class but that are for other markets (8540). It's a way to show the industry what your doing (blowing your own horn), without giving away Apple's precious secrets.
I am not guaranteeing that the e500 core is going to be the same core as the G5, but it sure looks like it could be. This core can have all kinds of APU's bolted on (which is part of the Book E standards) and they include Altivec. They could also drop (as someone suggested) things not necessary in the e500 core (serial int support, dual ethernet, can't remember what they suggested). That would only be necessary if these items took up any amount of die size (which I doubt most do).
As for the speeds listed, they are speeds specific to the 8540 product, not to any other product that may use the same core or a variant of it. Once the core (e500 in this case) is near completion, you begin to make chips for different sectors that utilize the core. Different performance levels, different applications. It is a modular design so you don't need to have the core 100% complete. In the same aspect you can be working on a G5 for Apple at the same time as it is a family class of processors that will all share similar aspects.
Now does all this "reasoning" (ha) imply that a G5 will be forthcoming this year? Heck no. So why do I think a G5 this year (and I'm thinking by July at the abosulte latest, but it may be earlier).
1. Wasn't Motorola's original time table for late '01 or early '02.
<a href="http://www.macuser.co.uk/MacSurfer/news/33343.html"; target="_blank">http://www.macuser.co.uk/MacSurfer/news/33343.html</a>;
<a href="http://news.zdnet.co.uk/story/0,,s2073809,00.html"; target="_blank">http://news.zdnet.co.uk/story/0,,s2073809,00.html</a>;
<a href="http://www.geek.com/procspec/apple/g5.htm"; target="_blank">http://www.geek.com/procspec/apple/g5.htm</a>;
2. Motorola is currently shipping products on their HiPerMOS7 (Seventh Generation High Performance Metal Oxide Semiconductor) .13 micron lithography and SOI (Silicon on Insulator) technology along with copper interconnects.
3. Motorola says the 8540 will sample in 2nd half of 2002. Since the 8540 was developed at Somerset (where Apples chips are developed) and Apple foots part of the bill there, I would expect to see an Apple G5 class processor ship before an embedded Motorola part.
4. If I lose that would cut into my Xmas 2002 in Vegas fund (what can I say, my family likes to party). I need those coins for the slots and those greenbacks for the tables.
EDIT------------------------------------------------------------------------------------------------
Just saw this take a look:
<a href="http://biz.yahoo.com/rf/020108/tp46588_1.html"; target="_blank">http://biz.yahoo.com/rf/020108/tp46588_1.html</a>;
[ 01-09-2002: Message edited by: Slacker ]</p>
<strong>If this weren't the case, you can be sure that IBM would have been fabbing G4s long ago (and we'd have broken the GHz barrier a long time ago. Sorry Moto, but it's true...</strong><hr></blockquote>
Yes, IBM really boosted their G3 to GHz+ speeds - NOT!
ibm WAS fabbing the g4. the didn't change the design though, just fabbing...heard that from an apple emp at wwdc.