<strong>Still, Dorsal said these G4s with DDR support were running at 1 GHz...same as current G4s.</strong><hr></blockquote>
nonono, s/he said like in china girl from bowie
oh well whatever...
Dorsal said: "They (7460) 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." again...
<strong>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.</strong><hr></blockquote>
Maybe because they don't have the DDR G4 and the mobo could at least be more efficient with the I/O (ethernet, ATA).
So if we put a DDR G4 in this mobo. We have what we wanted for a long time. and maybe this mobo support DDR PC-2700 @ 333 Mhz... who knows?
[quote] Maybe because they don't have the DDR G4 <hr></blockquote>
That's my point. There is no G4 that supports a DDR RAM frontside bus. If there were, then Apple could have used it in the XServe and forgot about spending the resources to develop a DDR hack solution.
This leads me to believe that there will be no true DDR support until the G5 and RapidIO arrive.
It's possible ( even likely) that Dorsal was testing prototype chips not made on a full scale fab. If those chips were designed to be fabbed on a 0.13 micron process (which he indeed suggests) then they won't be available 'till moto's 0.13 fabs are up and running (even if the chip itself is already finished), wich is just now happening, or is soon about to happen, depending on who you believe.
..so no JYD I don' buy your argument at all. That doesn't necessarily mean Dorsal is/was a real source though.
That's my point. There is no G4 that supports a DDR RAM frontside bus. If there were, then Apple could have used it in the XServe and forgot about spending the resources to develop a DDR hack solution.
This leads me to believe that there will be no true DDR support until the G5 and RapidIO arrive.</strong><hr></blockquote>
the technical aspects are above me but I don't think what Apple has done is a hack. All they did was make the memory bus DDR. They didn't really hack anything. They could have probably done that a while ago but instead were waiting on Motorola to do something with the FSB but couldn't wait any longer
<strong>That's my point. There is no G4 that supports a DDR RAM frontside bus. If there were, then Apple could have used it in the XServe and forgot about spending the resources to develop a DDR hack solution.
This leads me to believe that there will be no true DDR support until the G5 and RapidIO arrive.</strong><hr></blockquote>
The fact that Apple didn't use the supposed DDR-ready G4 in the Xserve simply means it wasn't ready or appropriate for the Xserve. It does mean that Apple has started buying lots of DDR RAM, however.
Given the statements by the Motorola rep about the positioning of MPX in the semiconductor market, however, I'm not optimistic about seeing a DDR MPX bus. Processor support for DDR will have to wait until the memory controller moves onto the chip. We currently have no hard information on when that will be. It could be MWNY, or it could be next year. It could be a modified G4, or a new fangled G5. That the Xserve didn't go there just means it was either not yet ready or not appropriate (i.e. the server market might prefer a tried and true processor, rather than a new unproven one).
First, I agree that it looks like DDR MPX will never happen. I also agree that that's not a bad thing.
[quote]Originally posted by Programmer:
<strong>That the Xserve didn't go there just means it was either not yet ready or not appropriate (i.e. the server market might prefer a tried and true processor, rather than a new unproven one).</strong><hr></blockquote>
Hell, for all I know the XServe customers are happy that no matter how much the CPUs get hammered, a certain amount of memory bandwidth for DMA is guaranteed.
I don't fully understand the function of the MPX Bus. Would anyone care to enlighten me, please? currently, I think it connects PCI-RAM-Disk I/O to the Processor, but I also think I'm incorrect. An ACSII diagram would be helpful.
<strong>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.</strong><hr></blockquote>
Given the statements by the Motorola rep about the positioning of MPX in the semiconductor market, however, I'm not optimistic about seeing a DDR MPX bus. Processor support for DDR will have to wait until the memory controller moves onto the chip.</strong><hr></blockquote>
It DOES make sense for moto to develop a DDR capable version of the next G4, if apple pays for the extra R&D.
Now, does it make sense for apple to pay? It might, if you consider that the chip is going to live on in the consumer macs (and portables, in the case of the 7460 version) for a long time after it's successor (what ever it may be) has taken over the PMs. In this light it doesn't appear to be such a stopgap solution.
Im not saying it's going to happen, just that I'm not so dead certain as some, about it not happening.
Because of the way they use the word [Boxes], I fear someone saw new the new case and/or Xserve type mobo, and assumed it would ship with the G5 CPU. I fear what is ready for July uses the G4 (if so hopefully in a quad configuration, -(sadly no entirely credible rumors of DDR FSB for G4's exist?))
If Steve said no G5 untill XXdate the whooshing sound you heard would be the price of AAPL stock going down the toilet.
If I bought '1.4' DDR dual machine with SDR or DDR in July (should such machine be available...) I'd consider myself hard done by if Apple release a 2.4 gig G5 in January.
That seems one helluva leap.
If they have 2.4 G5s by January...then why not ship some down clocked G5s now? If they are as powerful as rumours suggest then a 1 gig - 1.6 gig G5 on Rapid Io would suffice.
If all these rumours are true, Apple are playing one hell of a game of 'catch up'.
I think next Jan' is shaping up to be the time to buy...
MWSF will only be the time to buy if there is another "silent" Powermac update this summer. Otherwise, Apple is going to blow their load at MWNY and MWSF won't have much to offer.
We can safely say that Powermacs will get new muthaboards this summer. So the question is what sort of CPU? Either:
1. G5 w/ RapidIO, 130 nm process. PC 2100 or PC 2700.
2. G4 w/ RapidIO, 130 nm process. PC 2100 or PC 2700
3. G4 w/ 133 MHz MPX frontside bus, same as current G4 but clocked higher. XServe DDR hack solution.
One of these implementations is going to be introduced and Apple will milk it for a while. Unfortunately, #3 is most likely, with a 20-40% bump in clockspeed. #2 would be very nice, and probably would bring up to a 50% bump in clockspeed. #1 is a wet dream and could bring up to a 100% bump in clockspeed.
I fear #3 is what's going to happen. We're going to be stuck with a hacked DDR solution for years, and it may signal Apple's demise in professional markets that require high performance workstations.
It DOES make sense for moto to develop a DDR capable version of the next G4, if apple pays for the extra R&D.
Now, does it make sense for apple to pay? It might, if you consider that the chip is going to live on in the consumer macs (and portables, in the case of the 7460 version) for a long time after it's successor (what ever it may be) has taken over the PMs. In this light it doesn't appear to be such a stopgap solution.
Im not saying it's going to happen, just that I'm not so dead certain as some, about it not happening.</strong><hr></blockquote>
I disagree -- the better solution for consumer, desktop, and portable machines is to move to a RapidIO / on-chip memory controller solution. Better integration, easier to build motherboard, better performance. Apple's money would be better spent accelerating development of the forward looking solution, rather than developing a bus that only they will use and that makes it harder to build motherboards.
We can safely say that Powermacs will get new muthaboards this summer. So the question is what sort of CPU? Either:
1. G5 w/ RapidIO, 130 nm process. PC 2100 or PC 2700.
2. G4 w/ RapidIO, 130 nm process. PC 2100 or PC 2700
3. G4 w/ 133 MHz MPX frontside bus, same as current G4 but clocked higher. XServe DDR hack solution.
I fear #3 is what's going to happen. We're going to be stuck with a hacked DDR solution for years, and it may signal Apple's demise in professional markets that require high performance workstations.
i agree, #3 looks inevitable. otherwise, apple would have easily just released an Xserve without the DDR memory.
why go to the trouble with the DDR hack, if Rapid i/o is prime-time? They would have just stayed with PC133, unless they plan to push the fake DDR into the consumer level Imacs and the Powerbooks and give the Powermacs Rapid i/o + "G5" at mwny.
[quote] They would have just stayed with PC133, unless they plan to push the fake DDR into the consumer level Imacs and the Powerbooks and give the Powermacs Rapid i/o + "G5" at mwny.
<hr></blockquote>
Now there's an idea!
Since the G4 has a long life ahead of it in Apple's portables and low end desktops, then perhaps the DDR hack is going to live on in such computers. Meanwhile the Powermacs get G5s and RapidIO.
This could be interesting. Although I'm under the impression that a RapidIO unit could be "bolted" onto the G4 as easily as the MPX bus. If so then why not make all desktop PPC chips (G4 and G5) with RapidIO, and let Apple choose which kind of RAM they want to use? Apple could use PC2100 in the low end Macs, and PC2700 in the Powermacs.
Comments
<strong>Still, Dorsal said these G4s with DDR support were running at 1 GHz...same as current G4s.</strong><hr></blockquote>
nonono, s/he said like in china girl from bowie
oh well whatever...
Dorsal said: "They (7460) 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."
<strong>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.</strong><hr></blockquote>
Maybe because they don't have the DDR G4 and the mobo could at least be more efficient with the I/O (ethernet, ATA).
So if we put a DDR G4 in this mobo. We have what we wanted for a long time. and maybe this mobo support DDR PC-2700 @ 333 Mhz... who knows?
That's my point. There is no G4 that supports a DDR RAM frontside bus. If there were, then Apple could have used it in the XServe and forgot about spending the resources to develop a DDR hack solution.
This leads me to believe that there will be no true DDR support until the G5 and RapidIO arrive.
..so no JYD I don' buy your argument at all. That doesn't necessarily mean Dorsal is/was a real source though.
If he isn't, he's damn good though
<strong>
That's my point. There is no G4 that supports a DDR RAM frontside bus. If there were, then Apple could have used it in the XServe and forgot about spending the resources to develop a DDR hack solution.
This leads me to believe that there will be no true DDR support until the G5 and RapidIO arrive.</strong><hr></blockquote>
the technical aspects are above me but I don't think what Apple has done is a hack. All they did was make the memory bus DDR. They didn't really hack anything. They could have probably done that a while ago but instead were waiting on Motorola to do something with the FSB but couldn't wait any longer
<strong>That's my point. There is no G4 that supports a DDR RAM frontside bus. If there were, then Apple could have used it in the XServe and forgot about spending the resources to develop a DDR hack solution.
This leads me to believe that there will be no true DDR support until the G5 and RapidIO arrive.</strong><hr></blockquote>
The fact that Apple didn't use the supposed DDR-ready G4 in the Xserve simply means it wasn't ready or appropriate for the Xserve. It does mean that Apple has started buying lots of DDR RAM, however.
Given the statements by the Motorola rep about the positioning of MPX in the semiconductor market, however, I'm not optimistic about seeing a DDR MPX bus. Processor support for DDR will have to wait until the memory controller moves onto the chip. We currently have no hard information on when that will be. It could be MWNY, or it could be next year. It could be a modified G4, or a new fangled G5. That the Xserve didn't go there just means it was either not yet ready or not appropriate (i.e. the server market might prefer a tried and true processor, rather than a new unproven one).
[quote]Originally posted by Programmer:
<strong>That the Xserve didn't go there just means it was either not yet ready or not appropriate (i.e. the server market might prefer a tried and true processor, rather than a new unproven one).</strong><hr></blockquote>
Hell, for all I know the XServe customers are happy that no matter how much the CPUs get hammered, a certain amount of memory bandwidth for DMA is guaranteed.
<strong>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.</strong><hr></blockquote>
Ok, however. I am more interesting about G5.
I am very confused about G5 informations.
Looking at Geek, Architosh, The Register.
Architosh writes about G5 test boxes:
<a href="http://www.architosh.com/news/2002-02/2002b-0201-futurg5.phtml"; target="_blank">http://www.architosh.com/news/2002-02/2002b-0201-futurg5.phtml</a>;
The Register writes that the G5 are ready for volume production early next year with rates at 2.4GHz.
If this is true, than the G5 could ready in September with rates at 1GHz!?
Alsoft, i think writes in January about the G5, also Epson.
Steve Jobs knows, that everybody is waiting for the G5 and people are holding off orders.
If there is no G5 in the near future, why dosen´t he saying there is no G5?
<strong>
Given the statements by the Motorola rep about the positioning of MPX in the semiconductor market, however, I'm not optimistic about seeing a DDR MPX bus. Processor support for DDR will have to wait until the memory controller moves onto the chip.</strong><hr></blockquote>
It DOES make sense for moto to develop a DDR capable version of the next G4, if apple pays for the extra R&D.
Now, does it make sense for apple to pay? It might, if you consider that the chip is going to live on in the consumer macs (and portables, in the case of the 7460 version) for a long time after it's successor (what ever it may be) has taken over the PMs. In this light it doesn't appear to be such a stopgap solution.
Im not saying it's going to happen, just that I'm not so dead certain as some, about it not happening.
<a href="http://www.architosh.com/news/2002-02/2002b-0201-futurg5.phtml"; target="_blank">http://www.architosh.com/news/2002-02/2002b-0201-futurg5.phtml</a>;
Because of the way they use the word [Boxes], I fear someone saw new the new case and/or Xserve type mobo, and assumed it would ship with the G5 CPU. I fear what is ready for July uses the G4 (if so hopefully in a quad configuration, -(sadly no entirely credible rumors of DDR FSB for G4's exist?))
If Steve said no G5 untill XXdate the whooshing sound you heard would be the price of AAPL stock going down the toilet.
That seems one helluva leap.
If they have 2.4 G5s by January...then why not ship some down clocked G5s now? If they are as powerful as rumours suggest then a 1 gig - 1.6 gig G5 on Rapid Io would suffice.
If all these rumours are true, Apple are playing one hell of a game of 'catch up'.
I think next Jan' is shaping up to be the time to buy...
Lemon Bon Bon <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" />
We can safely say that Powermacs will get new muthaboards this summer. So the question is what sort of CPU? Either:
1. G5 w/ RapidIO, 130 nm process. PC 2100 or PC 2700.
2. G4 w/ RapidIO, 130 nm process. PC 2100 or PC 2700
3. G4 w/ 133 MHz MPX frontside bus, same as current G4 but clocked higher. XServe DDR hack solution.
One of these implementations is going to be introduced and Apple will milk it for a while. Unfortunately, #3 is most likely, with a 20-40% bump in clockspeed. #2 would be very nice, and probably would bring up to a 50% bump in clockspeed. #1 is a wet dream and could bring up to a 100% bump in clockspeed.
I fear #3 is what's going to happen. We're going to be stuck with a hacked DDR solution for years, and it may signal Apple's demise in professional markets that require high performance workstations.
[ 05-27-2002: Message edited by: Junkyard Dawg ]</p>
<strong>
It DOES make sense for moto to develop a DDR capable version of the next G4, if apple pays for the extra R&D.
Now, does it make sense for apple to pay? It might, if you consider that the chip is going to live on in the consumer macs (and portables, in the case of the 7460 version) for a long time after it's successor (what ever it may be) has taken over the PMs. In this light it doesn't appear to be such a stopgap solution.
Im not saying it's going to happen, just that I'm not so dead certain as some, about it not happening.</strong><hr></blockquote>
I disagree -- the better solution for consumer, desktop, and portable machines is to move to a RapidIO / on-chip memory controller solution. Better integration, easier to build motherboard, better performance. Apple's money would be better spent accelerating development of the forward looking solution, rather than developing a bus that only they will use and that makes it harder to build motherboards.
<strong>
We can safely say that Powermacs will get new muthaboards this summer. So the question is what sort of CPU? Either:
1. G5 w/ RapidIO, 130 nm process. PC 2100 or PC 2700.
2. G4 w/ RapidIO, 130 nm process. PC 2100 or PC 2700
3. G4 w/ 133 MHz MPX frontside bus, same as current G4 but clocked higher. XServe DDR hack solution.
I fear #3 is what's going to happen. We're going to be stuck with a hacked DDR solution for years, and it may signal Apple's demise in professional markets that require high performance workstations.
[ 05-27-2002: Message edited by: Junkyard Dawg ]</strong><hr></blockquote>
i agree, #3 looks inevitable. otherwise, apple would have easily just released an Xserve without the DDR memory.
why go to the trouble with the DDR hack, if Rapid i/o is prime-time? They would have just stayed with PC133, unless they plan to push the fake DDR into the consumer level Imacs and the Powerbooks and give the Powermacs Rapid i/o + "G5" at mwny.
<hr></blockquote>
Now there's an idea!
Since the G4 has a long life ahead of it in Apple's portables and low end desktops, then perhaps the DDR hack is going to live on in such computers. Meanwhile the Powermacs get G5s and RapidIO.
This could be interesting. Although I'm under the impression that a RapidIO unit could be "bolted" onto the G4 as easily as the MPX bus. If so then why not make all desktop PPC chips (G4 and G5) with RapidIO, and let Apple choose which kind of RAM they want to use? Apple could use PC2100 in the low end Macs, and PC2700 in the Powermacs.
I would hope that Apple has had enough time to leap that particular move.
Screed
G4 7470 1.5GHz supports PC2100 DDR-SDRAM @ 266MHZ <a href="http://www.applelinks.com/articles/2002/05/20020524162055.shtml"; target="_blank">http://www.applelinks.com/articles/2002/05/20020524162055.shtml</a>;
How credible is this? Does this guy check sources? Is this just the Register rumor circulating around?
If no G5 I want to believe this.
<img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" />