<strong>If this chip isn't comming until the second half of 2003. Why were the PM's updated with the odd DDR mod. Clearly the 970 will need a new logic board and will support DDR properly, and what is going to happen in January to make new machines OS X only?</strong><hr></blockquote>
I think the DDR mod will allow for a trickle down to the consumer models, so you have an iMac running a G4+++ with a true DDR bus, and a PowerMac running a 900 mhz bus with a PPC 970. Seems like a good concept to me.
Hopefully since IBM wants to use this chip for Linux systems as well as for Apple, and presumably interest other manufacturers in a CHRP type design, the CPU will come in a standard socket format, like the P4 or Athlon. Then CPU modules would not need to be custom made by Apple, and prices should go down quicker because the chips would be standardized. My $.02
<strong>Hopefully since IBM wants to use this chip for Linux systems as well as for Apple, and presumably interest other manufacturers in a CHRP type design, the CPU will come in a standard socket format, like the P4 or Athlon. Then CPU modules would not need to be custom made by Apple, and prices should go down quicker because the chips would be standardized. My $.02</strong><hr></blockquote>
I wouldn't count on this happening -- IBM and Apple make their money selling systems, not individual chips. Just ask Intel and AMD what the margin on chips is like!
There seems to be a conception that this new processor will be expensive, but I don't think that is necessarily so. Cost is generally a function of yield, and yield is generally a function of die size (at least in the first approximation). Since we now know that this will be a 52 million transistor chip on a 0.13 micron process, this processor should be in about the same ballpark as the 7455 (which is 0.18 microns and about 30 million) and less than the PIV (which is now 0.13 micron and over 60 million, I believe). IBM has traditionally charged more for its fabs, but they are some of the better fabs around so without detailed information its hard to say whether that will increase or decrease prices. The mere fact that this is a 64-bit processor doesn't affect pricing at all. The cost of motherboards to support a 900 MHz bus might be higher, but it'll probably be a narrower bus. So I don't see any evidence (either way) of significantly different costs. Apple could try to charge a premium for these new machines, except that they already charge a premium and they have some catching up to do in terms of the price/performance of their PowerMac line. Going back to single-processor machines would save on their processor costs significantly, giving them bigger margins and the ability to have a wide range of PowerMac configurations to offer and the full range of reasonable to outrageous prices.
I don't know why everyone's saying the 970 can't come in dual core configurations. I haven't seen any such assertion from the sources. Additionally, I don't know why MWSF '04 is the time-line people are giving for the introduction of the machines.
I think we all realize that Apple really needs to announce and ship these machines as soon as possible. All possible resources should be diverted to getting them out ASAP. As others have noted, if PM sales look dismal right now, they're going to look dreadful the longer these machines are delayed. Only individuals who enjoy throwing money away would purchase the current desktops. A year is far too long to wait. If Apple doesn't announce by April, it will be terribly unfortunate.
Apple had eleven years - between '84 and '95 - to decimate MS in the market place. Apple could have been in the position of dominance MS has been in for these many years, yet they failed to do so, and now every day of complacently in which Apple rests, it is sealing the fate of our platform. The coming fray that the 64 bit era will usher in is another golden opportunity - perhaps the last - for Apple to gain the marketshare it is rightfully entitled to. But that's not going to happen unless the company is bold, aggressive and committed to winning back the desktop.
Going back to single-processor machines would save on their processor costs significantly, giving them bigger margins and the ability to have a wide range of PowerMac configurations to offer and the full range of reasonable to outrageous prices.
it would also make for an easier motherboard layout, which would be helpful in getting them out sooner than later, and reducing mobo costs as well.
They had *better* not go back to single processor only configs--a lot of people are porting for multiprocessing for Apple, and one of the strengths of the PPC970 is that it supports SMP up to 16 processors. I don't think we'll see octos out of the gate, but if they drop duals completely people won't code for it, and it will just muddy everything.
When fully loaded, IBM Microelectronics' East Fishkill facility ? thus far known only as Building 323 ? will ramp up over the course of this year to full-scale commercial production by next January. The fab will rival IBM's sprawling 200-mm complex in Burlington, Vt., in terms of production capacity, a spokesman said. Indeed, the abundance of capacity has led IBM to position itself as a leading-edge foundry, soliciting business from the largest fabless semiconductor makers, including Analog Devices Inc. and Xilinx Inc.
Is the 970 first on tap?
The IBM press release on Monday Oct 14 only says 'next year'
<strong>They had *better* not go back to single processor only configs--a lot of people are porting for multiprocessing for Apple, and one of the strengths of the PPC970 is that it supports SMP up to 16 processors. I don't think we'll see octos out of the gate, but if they drop duals completely people won't code for it, and it will just muddy everything.</strong><hr></blockquote>I hope they do drop the duals. It was a poor strategy from the start, and just because they've had a poor strategy in the past doesn't mean they should continue with a poor strategy.
Duals are just not efficient for mainstream computers. They're great for niche-y situations, like people who render video all day, and should be an option for those who use many dual-optimized apps. But for most of us, they're just not worthwhile. A single fast processor makes much more sense.
<strong>If one 1.8GHz PPC970s isn't enough to clobber Intel and AMD, two should do the job. The 970 presumably has Altivec and a VERY high IPC. I wouldn't worry too much about the P4, although the Opteron could pose a threat.
Does anyone know if the 970 has separate or shared buses for multiple CPUs?</strong><hr></blockquote>
I don't know, the power4 is very thoughtfully designed. The x86, on the other hand, is not.
I bet a 1.8Ghz 970 can run with a 5Ghz Whatever-Intel-will-have-in-2003.
They should keep duals. Unless anyone has failed to notice the PM's come in at a workstation price, and for that price they ought to offer workstation performance. Duals are good: Good marketing angle, good OS to take advantage of it, good platform to enable it in the first place. The Mac platform is in the rare position to make duals good, because of the strict control Apple exerts.
Another thing that it does is enable a better spread of offerings throughout the low end. With all duals in the pro machines, Apple could serve up faster single CPU's in the consumer line-up without worrying about speed overlap. Apple needs to exploit advantages in hardware when/where-ever it can get them. Being marginally faster or on par with similarly priced windows hardware is nowhere near as good as being measurably faster at the same price points.
It all depends on how 'fast' the next gen PPC really is, but you can expect Intel to produce a quick answer.
Apple now has an OS which fully takes advantage of dual CPUs.
This CPU has a 6.4GB/s bus (compared to the G4+'s 1.3GB/s), so even if it is shared there will be enough bandwidth to run duals effectivly.
Double your CPU, double your power*.
As CPUs get smaller, cheaper and cooler, it will make even more sense to run multiple processors.
EETimes has an <a href="http://www.eet.com/semi/news/OEG20021014S0059"; target="_blank">article</a> up on MPF. In it they assert that this CPU has "what Motorola calls Altivec and Apple calls the Velocity Engine".
"The 970 also sports a cache-coherent, 900-MHz processor bus capable of data rates up to 6.4 Gbytes/second. It will support symmetric-multiprocessing configurations of up to 16 CPUs. That capability could be valuable for Apple's new line of Xserve entry-level servers."
A lot of the Power 4's amazing performance comes from the fact that it has 128 MB of L3 cache.... I'm postive this chip doesn't have anywhere near that..... so I would expect performance would be no where near comparable..... any word on how much cache these things will have?
<strong>Apple now has an OS which fully takes advantage of dual CPUs.
...
Double your CPU, double your power*.</strong><hr></blockquote>The first statement only applies if 1) the app is dual aware or 2) you're running two compute-intensive apps at once. Unless one of those two conditions is met, you might as well have one processor rather than two.
As for the second statement - do you work for Apple marketing?
Comments
<strong>
How much are Apple going to charge for the Power4 Lite 970 machine?
[ 10-14-2002: Message edited by: Addison ]</strong><hr></blockquote>
That partially depends on how much IBM charges for the chips.
So the question would be how much are these chips
gonna cost.
900mhz bus <img src="graemlins/smokin.gif" border="0" alt="[Chilling]" />
[ 10-14-2002: Message edited by: O and A ]</p>
<strong>If this chip isn't comming until the second half of 2003. Why were the PM's updated with the odd DDR mod. Clearly the 970 will need a new logic board and will support DDR properly, and what is going to happen in January to make new machines OS X only?</strong><hr></blockquote>
I think the DDR mod will allow for a trickle down to the consumer models, so you have an iMac running a G4+++ with a true DDR bus, and a PowerMac running a 900 mhz bus with a PPC 970. Seems like a good concept to me.
<a href="http://www.cnn.com/2002/TECH/biztech/10/14/ibm.apple.ap/index.html"; target="_blank">from cnn</a>
Now everyone is jumping on the "Apple will go IBM" bandwagon
<strong>CNN
Now everyone is jumping on the "Apple will go IBM" bandwagon
But will Apple jump on to that wagon?
<strong>Hopefully since IBM wants to use this chip for Linux systems as well as for Apple, and presumably interest other manufacturers in a CHRP type design, the CPU will come in a standard socket format, like the P4 or Athlon. Then CPU modules would not need to be custom made by Apple, and prices should go down quicker because the chips would be standardized. My $.02</strong><hr></blockquote>
I wouldn't count on this happening -- IBM and Apple make their money selling systems, not individual chips. Just ask Intel and AMD what the margin on chips is like!
There seems to be a conception that this new processor will be expensive, but I don't think that is necessarily so. Cost is generally a function of yield, and yield is generally a function of die size (at least in the first approximation). Since we now know that this will be a 52 million transistor chip on a 0.13 micron process, this processor should be in about the same ballpark as the 7455 (which is 0.18 microns and about 30 million) and less than the PIV (which is now 0.13 micron and over 60 million, I believe). IBM has traditionally charged more for its fabs, but they are some of the better fabs around so without detailed information its hard to say whether that will increase or decrease prices. The mere fact that this is a 64-bit processor doesn't affect pricing at all. The cost of motherboards to support a 900 MHz bus might be higher, but it'll probably be a narrower bus. So I don't see any evidence (either way) of significantly different costs. Apple could try to charge a premium for these new machines, except that they already charge a premium and they have some catching up to do in terms of the price/performance of their PowerMac line. Going back to single-processor machines would save on their processor costs significantly, giving them bigger margins and the ability to have a wide range of PowerMac configurations to offer and the full range of reasonable to outrageous prices.
I think we all realize that Apple really needs to announce and ship these machines as soon as possible. All possible resources should be diverted to getting them out ASAP. As others have noted, if PM sales look dismal right now, they're going to look dreadful the longer these machines are delayed. Only individuals who enjoy throwing money away would purchase the current desktops. A year is far too long to wait. If Apple doesn't announce by April, it will be terribly unfortunate.
Apple had eleven years - between '84 and '95 - to decimate MS in the market place. Apple could have been in the position of dominance MS has been in for these many years, yet they failed to do so, and now every day of complacently in which Apple rests, it is sealing the fate of our platform. The coming fray that the 64 bit era will usher in is another golden opportunity - perhaps the last - for Apple to gain the marketshare it is rightfully entitled to. But that's not going to happen unless the company is bold, aggressive and committed to winning back the desktop.
Going back to single-processor machines would save on their processor costs significantly, giving them bigger margins and the ability to have a wide range of PowerMac configurations to offer and the full range of reasonable to outrageous prices.
it would also make for an easier motherboard layout, which would be helpful in getting them out sooner than later, and reducing mobo costs as well.
neye
i believe that is because the press release from IBM says they are single cored.
This from July: (Article is about 3000mm wafers)
When fully loaded, IBM Microelectronics' East Fishkill facility ? thus far known only as Building 323 ? will ramp up over the course of this year to full-scale commercial production by next January. The fab will rival IBM's sprawling 200-mm complex in Burlington, Vt., in terms of production capacity, a spokesman said. Indeed, the abundance of capacity has led IBM to position itself as a leading-edge foundry, soliciting business from the largest fabless semiconductor makers, including Analog Devices Inc. and Xilinx Inc.
Is the 970 first on tap?
The IBM press release on Monday Oct 14 only says 'next year'
<a href="http://www.siliconstrategies.com/story/OEG20020719S0037"; target="_blank">http://www.siliconstrategies.com/story/OEG20020719S0037</a>;
<strong>They had *better* not go back to single processor only configs--a lot of people are porting for multiprocessing for Apple, and one of the strengths of the PPC970 is that it supports SMP up to 16 processors. I don't think we'll see octos out of the gate, but if they drop duals completely people won't code for it, and it will just muddy everything.</strong><hr></blockquote>I hope they do drop the duals. It was a poor strategy from the start, and just because they've had a poor strategy in the past doesn't mean they should continue with a poor strategy.
Duals are just not efficient for mainstream computers. They're great for niche-y situations, like people who render video all day, and should be an option for those who use many dual-optimized apps. But for most of us, they're just not worthwhile. A single fast processor makes much more sense.
<strong>If one 1.8GHz PPC970s isn't enough to clobber Intel and AMD, two should do the job. The 970 presumably has Altivec and a VERY high IPC. I wouldn't worry too much about the P4, although the Opteron could pose a threat.
Does anyone know if the 970 has separate or shared buses for multiple CPUs?</strong><hr></blockquote>
I don't know, the power4 is very thoughtfully designed. The x86, on the other hand, is not.
I bet a 1.8Ghz 970 can run with a 5Ghz Whatever-Intel-will-have-in-2003.
Another thing that it does is enable a better spread of offerings throughout the low end. With all duals in the pro machines, Apple could serve up faster single CPU's in the consumer line-up without worrying about speed overlap. Apple needs to exploit advantages in hardware when/where-ever it can get them. Being marginally faster or on par with similarly priced windows hardware is nowhere near as good as being measurably faster at the same price points.
It all depends on how 'fast' the next gen PPC really is, but you can expect Intel to produce a quick answer.
Apple now has an OS which fully takes advantage of dual CPUs.
This CPU has a 6.4GB/s bus (compared to the G4+'s 1.3GB/s), so even if it is shared there will be enough bandwidth to run duals effectivly.
Double your CPU, double your power*.
As CPUs get smaller, cheaper and cooler, it will make even more sense to run multiple processors.
EETimes has an <a href="http://www.eet.com/semi/news/OEG20021014S0059"; target="_blank">article</a> up on MPF. In it they assert that this CPU has "what Motorola calls Altivec and Apple calls the Velocity Engine".
"The 970 also sports a cache-coherent, 900-MHz processor bus capable of data rates up to 6.4 Gbytes/second. It will support symmetric-multiprocessing configurations of up to 16 CPUs. That capability could be valuable for Apple's new line of Xserve entry-level servers."
w00tz0r 16-CPU workstations
Barto
*Excuse the pun.
<strong>Apple now has an OS which fully takes advantage of dual CPUs.
...
Double your CPU, double your power*.</strong><hr></blockquote>The first statement only applies if 1) the app is dual aware or 2) you're running two compute-intensive apps at once. Unless one of those two conditions is met, you might as well have one processor rather than two.
As for the second statement - do you work for Apple marketing?