OK, now we have some more evidence to base the speculation on.
The new iMac is indeed a G4.
Which means, the G5 is definitely on its way.
I remember posting in one of these threads a while ago reminding people that for Apple to have the impact of the original iMac, they must push its specs very close to the highest end PowerMac, just like they did in 1998.
Of course, this doesn't mean they won't re-introduce a clearer differentiation between the professional and consumer lines with a G5 at the next MacWorld.
God knows they seriously need G5s that perform at least twice as fast as the current G4s in their PowerMacs but whether we'll get them tomorrow, I don't know.
<strong>It's been a while and much has happened in the interim, but some things never change, especially the over optimistic and over pessimistic fellings shared on this board. But it's good to know some things never change. One thing i don't know is how this rumor about the G5 is so widespread. So far I have yet to see one in any lab allthough I don't doubt they exist; inside Apple labs. What we have seen though, is widespread use of the 7460 which is basically a 7450 with an improved method of manufacter. Still this is nothing to sneeze at. They offer linear performance over the 7450 at better speeds. The range so far has been ~900-1400MHz but it is hard to give exact numbers due to the variety of motherboards they exist on. I was shocked when the newer motherboard we were working on were not released at the Expo this past summer. Fully working DDR-SDRAM motherboards were ready with a full assortment of modern motherboard features including Fibre Firewire at speeds up to 1600Mbps with a fibre port and 2 lower speed (800Mb) normal ports. USB1.1 was still there but the board had support for DDR-SDRAM and an advanced system bus running at 266MHz. They were to include CPU's running at up to 1GHz. Perhaps faster CPU's were hard to come by.
Of couse they could have put that plan on ice and wait for the recently announced 333Mhz DDR-SDRAM. The board was fully compatible with the newer SDRAM standard and easy to impliment. this of course would imply the cpu bus to the main controller would be accordingly sped up to provice sufficient bandwidth. Internally on the main controller (memory+PCI+peripherals (there is no seperate southbridge controller)) there is a hyper transport link from PCI controller and peripherals such as ATA/133, USB, audio (also new), etc. Firewire and ethernet have their own seperate connections. This is an advanced peice of silicon. No, to me the only reason to release this board that is all ready this Expo is simple; many of the advanced features would go unused. perhaps there would have been a lack of advanced firewire peripherals. Or maybe they wanted to see the outcome of the memory wars between RAMBUS and the DDR consortium (there was in fact RAMBUS based prototypes of G4 systems floating around that we never came in contact with). In just over a month we will see a leap in performance from Apple's high end. It should be enough to justify their role in high end applications for years to come. We'll see.</strong><hr></blockquote>
Moving this up to the top. Where the hell are you Dorsal? Weed need the insider jam baby!
Hm, I would not expect too much of a company like Moto in the current state.
However, I wouldn't expect much from any company like Moto that does not even show a major product like the PowerPC processor in the featured products list on its own website.
<strong>However, I wouldn't expect much from any company like Moto that does not even show a major product like the PowerPC processor in the featured products list on its own website.</strong><hr></blockquote>
That's hardly a fair accusation -- go to motorola.com, click semiconductors, click PowerPC ISA. Why aren't they on the main page, front and center? Perhaps because the main visitors to Motorola's main page are looking for consumer products. If you're looking for chips then you'd probably go straight to e-www.motorola.com, which does mention PowerPC right there.
Go to IBM's main page, they don't even mention their Microelectronics division. Motorola and IBM are like this because they are huge, diverse companies with lots of varied products.
I'm just a novice when talking about procs, but I was wondering....could the G5 be sharing the same technologies that the PS3's (Playstation 3) chip is supposedly going to have? After all, from outsider's articles he posted, the location of the new G5 with the book e core seems to be in the same place as the PS3's chip, I think codenamed the "cell"
<strong>Here's some older but obscure links on the e500 core (core of the G5).</strong><hr></blockquote>
Everybody keeps talking about the e500 core like it is definitely going to be the core of the G5 that Apple will use. There is no proof of this, and it seems to arise soley from the fact that Motorola has chosen to announce this core as the centerpiece of their G5 embedded strategy. The e500 core is less superscalar than the G4's core, and clearly not intended to be a desktop processor. None of their documentation even hints at the core being intended for the desktop market (just networking, automotive and embedded consumer devices). Very low power is the emphasis.
The PowerPC design center at Somerset has done multiple cores at the same time before -- 3, in fact. The 603, 604, and 620 all were designed in the same time frame, and they are all the "G2" PowerPCs. The 603 carried on to become the 740/750 (G3), while the 604 was dropped because the 750 turned out so well the extra expense of the 604 wasn't enough of a payoff. The 7400 (G4) was a 750 with AltiVec added and several improvements to improve performance that were along the lines of things that had been in the 604. The 620 never got established in its target market (servers) and was dropped due to a total lack of demand.
None of this implies that the e500 core must be the core of the G5 that Apple will use. That Apple's G5 will likely be 64-bit implies that it is not the same core. Indeed, the long time between the 7400 and the 7450, and now the 7460 seems to imply one of three things: (a) the design center was massively downsized, (b) they're all just loafing about, or (c) they're working on something else at the same time.
<strong>I'm just a novice when talking about procs, but I was wondering....could the G5 be sharing the same technologies that the PS3's (Playstation 3) chip is supposedly going to have? After all, from outsider's articles he posted, the location of the new G5 with the book e core seems to be in the same place as the PS3's chip, I think codenamed the "cell"</strong><hr></blockquote>
I doubt very much that the two are at all related. Its possible that Sony could abandon the MIPS processor that it currently uses, but I doubt it (although I suppose there is always hope). I didn't see any PS3 references on either of the links Outsider posted -- is there one in particular that you are looking at?
You're right, the e500 core could be a very scaled down version of the core that WILL be used in G5 desktop computers. Hopefully Apple has had a heavy hand in the design of the processors that will be going into their future computers and will just use Motorola as a manufacturer. Or better yet have a manufacturer with a better track record make an Apple designed processor with technology licensed from Motorola. The problem is that none of these companies are disclosing ANY information... it's like they're deliberately trying to close down the FH forums! Those bastards!
This post of Dorsal's suggests that he's a fake. If he had actually been testing G4 mobos with a frontside bus that supports DDR RAM at 266 MHz, then we would have seen these on the XServe. Instead we find out that the G4 doesn't even support a bus like the one Dorsal claims to have tested.
<strong>This post of Dorsal's suggests that he's a fake. If he had actually been testing G4 mobos with a frontside bus that supports DDR RAM at 266 MHz, then we would have seen these on the XServe. Instead we find out that the G4 doesn't even support a bus like the one Dorsal claims to have tested.
It's a hoax, people.</strong><hr></blockquote>
The 7460/ 7470 already supports 266MHz/ 333MHz DDR-RAM.
<strong>This post of Dorsal's suggests that he's a fake. If he had actually been testing G4 mobos with a frontside bus that supports DDR RAM at 266 MHz, then we would have seen these on the XServe. Instead we find out that the G4 doesn't even support a bus like the one Dorsal claims to have tested.
It's a hoax, people.</strong><hr></blockquote>
! Not A Hoax !
-at least some info too accurate. And now that I think about it, this makes me more hopefull of a FSB that supports DDR in July. A DDR disparity between the Xserve and Powermac would be very Apple. But I think Apple is in part trying to make up for a slow FSB by offloading some tasks traditionally performed by PC CPU's, thereby utilizing their modular design to speed things up.
P.S. Does anyone have more info on what tasks bypass the G4?
And what about rates of silicon errata being not as harsh in G4? or does that belong in new thread?
According to rumor, maybe even a rumor that started with Dorsal's post.
Dorsal posted this last year. If such Macs were being field tested so long ago, then why doesn't the XServe have a bus that supports 266 MHz DDR-RAM?
The only explanation is that Dorsal was not actually testing such hardware. Otherwise the XServe would have used it.
This doesn't mean that the G5 doesn't exist, or that it won't be out soon....all it means is that Dorsal is a hoax.</strong><hr></blockquote>
The 7470 isn´t available now. The 7455 don´t have full DDR-RAM support.
Moto are manufacture now chips in 0.13µm. A prototype of the 7470 exists more than a year ago, i think.
Another thought: Moto are now fab chips in 0.13µm. This means that the G5 (MPC8500) could be fab, too. The e500 core already exists, the MPC8540 is ready for shipping.
That geeknews report is nonsense. Also, why would the G5 be cancelled, and yet they have highly detailed specs on the G5! That doesn't make any sense.
What I'm saying is that if a G4 supported a DDR frontside bus, and that G4 was about to ship, then why would Apple use a DDR hack on the XServe? It would make more sense to add true DDR support to the XServe using the CPU.
True, it could be that Apple wanted the XServe out in time for edu buying, and they can ship the Powermacs much later, even September.
Still, Dorsal said these G4s with DDR support were running at 1 GHz...same as current G4s.
Comments
<strong>haha. So what are you saying then Programmer?</strong><hr></blockquote>
Nothing, I thought I made that clear...
The new iMac is indeed a G4.
Which means, the G5 is definitely on its way.
I remember posting in one of these threads a while ago reminding people that for Apple to have the impact of the original iMac, they must push its specs very close to the highest end PowerMac, just like they did in 1998.
Of course, this doesn't mean they won't re-introduce a clearer differentiation between the professional and consumer lines with a G5 at the next MacWorld.
God knows they seriously need G5s that perform at least twice as fast as the current G4s in their PowerMacs but whether we'll get them tomorrow, I don't know.
<strong>It's been a while and much has happened in the interim, but some things never change, especially the over optimistic and over pessimistic fellings shared on this board. But it's good to know some things never change. One thing i don't know is how this rumor about the G5 is so widespread. So far I have yet to see one in any lab allthough I don't doubt they exist; inside Apple labs. What we have seen though, is widespread use of the 7460 which is basically a 7450 with an improved method of manufacter. Still this is nothing to sneeze at. They offer linear performance over the 7450 at better speeds. The range so far has been ~900-1400MHz but it is hard to give exact numbers due to the variety of motherboards they exist on. I was shocked when the newer motherboard we were working on were not released at the Expo this past summer. Fully working DDR-SDRAM motherboards were ready with a full assortment of modern motherboard features including Fibre Firewire at speeds up to 1600Mbps with a fibre port and 2 lower speed (800Mb) normal ports. USB1.1 was still there but the board had support for DDR-SDRAM and an advanced system bus running at 266MHz. They were to include CPU's running at up to 1GHz. Perhaps faster CPU's were hard to come by.
Of couse they could have put that plan on ice and wait for the recently announced 333Mhz DDR-SDRAM. The board was fully compatible with the newer SDRAM standard and easy to impliment. this of course would imply the cpu bus to the main controller would be accordingly sped up to provice sufficient bandwidth. Internally on the main controller (memory+PCI+peripherals (there is no seperate southbridge controller)) there is a hyper transport link from PCI controller and peripherals such as ATA/133, USB, audio (also new), etc. Firewire and ethernet have their own seperate connections. This is an advanced peice of silicon. No, to me the only reason to release this board that is all ready this Expo is simple; many of the advanced features would go unused. perhaps there would have been a lack of advanced firewire peripherals. Or maybe they wanted to see the outcome of the memory wars between RAMBUS and the DDR consortium (there was in fact RAMBUS based prototypes of G4 systems floating around that we never came in contact with). In just over a month we will see a leap in performance from Apple's high end. It should be enough to justify their role in high end applications for years to come. We'll see.</strong><hr></blockquote>
Moving this up to the top. Where the hell are you Dorsal? Weed need the insider jam baby!
However, I wouldn't expect much from any company like Moto that does not even show a major product like the PowerPC processor in the featured products list on its own website.
Sad but true.
Moto's sig: We don't link to PPC.
<strong>
Moto's sig: We don't link to PPC.
Maybe the sig can change to this:
We admit we suck
<strong>However, I wouldn't expect much from any company like Moto that does not even show a major product like the PowerPC processor in the featured products list on its own website.</strong><hr></blockquote>
That's hardly a fair accusation -- go to motorola.com, click semiconductors, click PowerPC ISA. Why aren't they on the main page, front and center? Perhaps because the main visitors to Motorola's main page are looking for consumer products. If you're looking for chips then you'd probably go straight to e-www.motorola.com, which does mention PowerPC right there.
Go to IBM's main page, they don't even mention their Microelectronics division. Motorola and IBM are like this because they are huge, diverse companies with lots of varied products.
Dorsal man where the hell are you?
This one says products expected to be available in the first half of 2002.
<a href="http://www.improvsys.com/Dynamic/techonline071301.html"; target="_blank">http://www.improvsys.com/Dynamic/techonline071301.html</a>;
This one is the only concrete mention of a 7 stage pipeline (like the one in the 7450) being used in the e500.
<a href="http://www.siliconstrategies.com/story/OEG20010612S0064"; target="_blank">http://www.siliconstrategies.com/story/OEG20010612S0064</a>;
[ 01-19-2002: Message edited by: Outsider ]</p>
<strong>Here's some older but obscure links on the e500 core (core of the G5).</strong><hr></blockquote>
Everybody keeps talking about the e500 core like it is definitely going to be the core of the G5 that Apple will use. There is no proof of this, and it seems to arise soley from the fact that Motorola has chosen to announce this core as the centerpiece of their G5 embedded strategy. The e500 core is less superscalar than the G4's core, and clearly not intended to be a desktop processor. None of their documentation even hints at the core being intended for the desktop market (just networking, automotive and embedded consumer devices). Very low power is the emphasis.
The PowerPC design center at Somerset has done multiple cores at the same time before -- 3, in fact. The 603, 604, and 620 all were designed in the same time frame, and they are all the "G2" PowerPCs. The 603 carried on to become the 740/750 (G3), while the 604 was dropped because the 750 turned out so well the extra expense of the 604 wasn't enough of a payoff. The 7400 (G4) was a 750 with AltiVec added and several improvements to improve performance that were along the lines of things that had been in the 604. The 620 never got established in its target market (servers) and was dropped due to a total lack of demand.
None of this implies that the e500 core must be the core of the G5 that Apple will use. That Apple's G5 will likely be 64-bit implies that it is not the same core. Indeed, the long time between the 7400 and the 7450, and now the 7460 seems to imply one of three things: (a) the design center was massively downsized, (b) they're all just loafing about, or (c) they're working on something else at the same time.
<strong>I'm just a novice when talking about procs, but I was wondering....could the G5 be sharing the same technologies that the PS3's (Playstation 3) chip is supposedly going to have? After all, from outsider's articles he posted, the location of the new G5 with the book e core seems to be in the same place as the PS3's chip, I think codenamed the "cell"</strong><hr></blockquote>
I doubt very much that the two are at all related. Its possible that Sony could abandon the MIPS processor that it currently uses, but I doubt it (although I suppose there is always hope). I didn't see any PS3 references on either of the links Outsider posted -- is there one in particular that you are looking at?
It's a hoax, people.
<strong>This post of Dorsal's suggests that he's a fake. If he had actually been testing G4 mobos with a frontside bus that supports DDR RAM at 266 MHz, then we would have seen these on the XServe. Instead we find out that the G4 doesn't even support a bus like the one Dorsal claims to have tested.
It's a hoax, people.</strong><hr></blockquote>
The 7460/ 7470 already supports 266MHz/ 333MHz DDR-RAM.
<strong>This post of Dorsal's suggests that he's a fake. If he had actually been testing G4 mobos with a frontside bus that supports DDR RAM at 266 MHz, then we would have seen these on the XServe. Instead we find out that the G4 doesn't even support a bus like the one Dorsal claims to have tested.
It's a hoax, people.</strong><hr></blockquote>
! Not A Hoax !
-at least some info too accurate. And now that I think about it, this makes me more hopefull of a FSB that supports DDR in July. A DDR disparity between the Xserve and Powermac would be very Apple. But I think Apple is in part trying to make up for a slow FSB by offloading some tasks traditionally performed by PC CPU's, thereby utilizing their modular design to speed things up.
P.S. Does anyone have more info on what tasks bypass the G4?
And what about rates of silicon errata being not as harsh in G4? or does that belong in new thread?
[ 05-27-2002: Message edited by: pey/coy-ote ]</p>
The 7460/ 7470 already supports 266MHz/ 333MHz DDR-RAM.
<hr></blockquote>
According to rumor, maybe even a rumor that started with Dorsal's post.
Dorsal posted this last year. If such Macs were being field tested so long ago, then why doesn't the XServe have a bus that supports 266 MHz DDR-RAM?
The only explanation is that Dorsal was not actually testing such hardware. Otherwise the XServe would have used it.
This doesn't mean that the G5 doesn't exist, or that it won't be out soon....all it means is that Dorsal is a hoax.
<strong>
According to rumor, maybe even a rumor that started with Dorsal's post.
Dorsal posted this last year. If such Macs were being field tested so long ago, then why doesn't the XServe have a bus that supports 266 MHz DDR-RAM?
The only explanation is that Dorsal was not actually testing such hardware. Otherwise the XServe would have used it.
This doesn't mean that the G5 doesn't exist, or that it won't be out soon....all it means is that Dorsal is a hoax.</strong><hr></blockquote>
The 7470 isn´t available now. The 7455 don´t have full DDR-RAM support.
Moto are manufacture now chips in 0.13µm. A prototype of the 7470 exists more than a year ago, i think.
Another thought: Moto are now fab chips in 0.13µm. This means that the G5 (MPC8500) could be fab, too. The e500 core already exists, the MPC8540 is ready for shipping.
Have you read this?
<a href="http://www.geek.com/news/geeknews/2002may/bch20020524011893.htm"; target="_blank">http://www.geek.com/news/geeknews/2002may/bch20020524011893.htm</a>;
If IBM fab the next generation PowerPC chip for Apple we can wait til MWNY 2003, i think.
What I'm saying is that if a G4 supported a DDR frontside bus, and that G4 was about to ship, then why would Apple use a DDR hack on the XServe? It would make more sense to add true DDR support to the XServe using the CPU.
True, it could be that Apple wanted the XServe out in time for edu buying, and they can ship the Powermacs much later, even September.
Still, Dorsal said these G4s with DDR support were running at 1 GHz...same as current G4s.