My point is that it doesn't need to be. Not every machine has to have the highest performance. I'm not saying that the iBook, iMac, and PowerBook won't move up the performance "food chain" -- just that they won't suddenly leap to the top of it.
The "MHz myth" is going to get even more significant with this new PowerPC chip. Intel is going to be up to 3 or 4 GHz and it sounds like IBM is aiming for about 2 GHz. The efficiency per clock cycle will be much better, however, which means that the importance of the "MHz myth" will be even greater. Get used to it. The concept is valid and correct, and MHz are only going to become more meaningless in determining processor performance as time goes on with the coming of multi-core, multi-threaded, SIMD, and clockless designs. Benchmarks need to catch up with reality at some point.</strong><hr></blockquote>
Actually, I was under the assumption that Intel was hearing a lot of critisim that the Plll was actaully a better chip at lower speeds then the Pllll and that Intel's next P5 or whatever was going to come in at much lower speeds. If so, then the mhz myth may not be an issue unless AMD tries to play off it.
Uhmmm, actually they do. IBM's research labs have announced several advances that are capable of making each transistor "do more" -- or more accurately, do the same with less power.
</strong><hr></blockquote>
Not the same thing... they will still require high feature counts to implement functionality and that is going to result in big chips (and high power consumption at the targeted 2 GHz speed). Yes it will be efficient for a huge 2 GHz chip, but it will still be aimed at desktops and low-end servers.
[quote]<strong>
Nonsense. The ISA is why the P4 is as complex as it is, not the capabiliy of the chip. Intel engineers made several tradeoffs for clock speed, power consumption being just one of them. There is no reason to suppose that IBM had to make the same compromises in their design. For all you know, this chip runs cooler or as cool as the MPC7455. (OK, probably not, but the point is that we don't know and won't know until IBM releases it's thermal profile.)
</strong><hr></blockquote>
Actually from what I've read & been told the x86 decoding isn't really that much of the chip. Consider that the P-II, P-III, K7 have all done x86 instruction decoding quite effectively on much smaller transistor budgets. The cost of this in the PentiumIV didn't increase -- instead Intel threw more transistors at longer pipelines, bigger caches, more internal tables and registers, etc. As chips get bigger the fraction of the chip required to decode the x86 instructions goes down. And if you look at the POWER4 design you'll see that IBM is decoding (or cracking) their ISA in a similar manner, with similar associated costs in terms of circuitry.
I'd guess that IBM is aiming somewhere in the 60-100 million transistor range, and even on the better and considerably smaller process they are likely to use this will result in a hotter chip than the 7455... and we've yet to see Moto produce a 0.13 (or smaller) 74xx which will run much cooler if not faster.
The original statement I objected to said that Apple should move their entire lineup to the new IBM chip, if only to symbolically dump Motorola. This is misguided -- Motorola decided that aiming their top PowerPC at the embedded market was a better business move than aiming it at the top of the desktop market. Given that they have only one desktop customer, and that that customer is less than 5% of the desktop market this makes a good deal of sense, even though it is unfortunate for Apple's desktop machines. With a market of that size it just isn't worth Motorola's investment to try and compete with Intel and AMD. The embedded purposed designs are good for portables and other low-power requirements though.
IBM, on the other hand, has its own uses for desktop chips and Apple therefore becomes a way to dramatically increase the size of the market for their chips. They have already sunk a huge investment into POWER4 for their very high end machines so now it only makes sense to leverage that investment into other markets in order to recoup the original R&D cost. Now that Apple has a solid Unix-based OS I also think that their is an opportunity for IBM to broaden the appeal of their machines beyond the AIX & Linux markets, assuming that they can convince Apple that it is also in Apple's best interests (which it is). Motorola just isn't the right partner for Apple in their high end -- IBM is. Having multiple partners was the whole reason for AIM in the first place, so it is good to see that they appear to be getting back on track. Things are looking a lot more promising than they have in the last couple of years.
<strong>The original statement I objected to said that Apple should move their entire lineup to the new IBM chip, if only to symbolically dump Motorola. This is misguided</strong><hr></blockquote>
I see. And yes, you're correct of course. I'd be insane to make product decisions on any other basis than what makes the most business sense.
[quote]<strong>With a market of that size it just isn't worth Motorola's investment to try and compete with Intel and AMD. The embedded purposed designs are good for portables and other low-power requirements though.</strong><hr></blockquote>
Again I agree although I would just caution that the PPC line does a good job of blurring the distinction between desktop and laptop when it comes to thermal characteristics so thinking in terms of "embedded vs desktop" may put you into a box that doesn't really exist. IBM is very power conscious in their designs and they sell to the embedded market also so they may well have had an eye to the high end of the embedded market as they scaled down the Power4 design. This chip could be in laptops 6 mos after the first version ships. (Maybe.)
[quote]<strong>Things are looking a lot more promising than they have in the last couple of years.</strong><hr></blockquote>
Absolutely. Moto is not going away and IBM is coming out with new products for the Mac. I haven't been this encouraged since the PPC first beat the P5. As someone once said, between IBM and Motorola the roadmap looks very interesting.
Of course Apple won't move their entire lineup to the IBM PPC 64...but it would be damn cool if they did. I don't think anyone would complain that their iMac is faster than a 4 GHz Pentium IV, but hey, I've been wrong before. Maybe some Mac users would whine about their Macs being too fast?
As for heat and power consumption, fsck all that! Who gives a damn if an iMac uses an extra 30 watts? It's NOTHING compared to the energy wasted by driving SUVs...and most people seem to LOVE making up for small dicks with giant SUVs. There's no reason Apple couldn't use quality fans in the iMac and laptops so that they don't sound like jet engines. In fact the current iMac's fan is barely audible. Drop a 2 GHz Power 5 in that iMac and then MAYBE we might actually be able to hear the fan! Oh my!
<strong>As for heat and power consumption, fsck all that! Who gives a damn if an iMac uses an extra 30 watts?</strong><hr></blockquote>
Apple's industrial designers do. An extra 30 watts from the CPU would mean three times more heat to dissipate, which makes it much harder to stuff it into tight spaces like the iMac, the eMac and the PowerBook.
It'll trickle down to those machines after a die shrink or two. In the meantime, the G4 seems to be scaling nicely. At last. A move to an onboard memory controller and RapidIO would give it a new lease on life.
In the meantime, the G4 seems to be scaling nicely. At last.</strong><hr></blockquote>
It does? In February (actually 1/28) they shipped 1 Gig machines. Six months later they announced 1.25 Gig machines but I have yet to see one. Increasing 25% every six months is only a little worse than Moore's Law, but they don't even seem to be able to do that.
<strong>Again I agree although I would just caution that the PPC line does a good job of blurring the distinction between desktop and laptop when it comes to thermal characteristics so thinking in terms of "embedded vs desktop" may put you into a box that doesn't really exist. IBM is very power conscious in their designs and they sell to the embedded market also so they may well have had an eye to the high end of the embedded market as they scaled down the Power4 design. This chip could be in laptops 6 mos after the first version ships. (Maybe.)
</strong><hr></blockquote>
We haven't seen what a no-holds-barred desktop PowerPC can do since the 604e & 620. The best situation would be for IBM to target the limit of what next year's process can do, allowing Apple to build the fastest possible PowerPC at the time. It will take (slightly) longer to bring it down the line, but it jumps as far ahead as possible and gives them more time for the next iteration.
We haven't seen what a no-holds-barred desktop PowerPC can do since the 604e & 620. The best situation would be for IBM to target the limit of what next year's process can do, allowing Apple to build the fastest possible PowerPC at the time. It will take (slightly) longer to bring it down the line, but it jumps as far ahead as possible and gives them more time for the next iteration.
ha! this should have been done last year! so how long to wait? again a year ? or two? ... ok - IF it's right an Mot will use .09 and RapidIO we have the time we need. but with the current G4's at .18 and still NOT at 1.25Ghz... we don't have the time. apple can't stay at 2*1Ghz for 7 months and at 2*1.25 Ghz another 7 months... if that happens, i'd be very dissapointed.
see: i'm a freak - i wanted to buy a new computer at the time it delivers 3x performance in every area of my previous machine. with the change from the 9500/160Mhz 604e and 96MB Ram to my G4/400 i did a good jump... but now. 100Mhz bus to 166Mhz bus speed. ATA66 to ATA100 etc... i don't see that it's worth it... so. i've to wait at least another 7-8 months to buy a new killer-machine (and remember: i don't buy the fastest and most expensive machine but the one with the best price/performance (mostly the cheapest PowerMac)
We haven't seen what a no-holds-barred desktop PowerPC can do since the 604e & 620. The best situation would be for IBM to target the limit of what next year's process can do, allowing Apple to build the fastest possible PowerPC at the time. It will take (slightly) longer to bring it down the line, but it jumps as far ahead as possible and gives them more time for the next iteration.
We havent seen an equivelent to the 604e becouse Apple killed the clones, and as a result IBM stoped development of the CHRP. Now that Linex is maturing, and IBM is investing lots of $ in it, we might see the rebirth of the CHRP, and as a result more computers bieng sold that use PPC chips. This could only bring more revinue for IBM and Motorolla as the PPC becomes a more viable competator to the Intel/AMD offerings. Apple will be able to take advantage of it, but I doubt it will be due to number of computers that Apple sells, or for that matter Apple and IBM combined.
Comments
<strong>
My point is that it doesn't need to be. Not every machine has to have the highest performance. I'm not saying that the iBook, iMac, and PowerBook won't move up the performance "food chain" -- just that they won't suddenly leap to the top of it.
The "MHz myth" is going to get even more significant with this new PowerPC chip. Intel is going to be up to 3 or 4 GHz and it sounds like IBM is aiming for about 2 GHz. The efficiency per clock cycle will be much better, however, which means that the importance of the "MHz myth" will be even greater. Get used to it. The concept is valid and correct, and MHz are only going to become more meaningless in determining processor performance as time goes on with the coming of multi-core, multi-threaded, SIMD, and clockless designs. Benchmarks need to catch up with reality at some point.</strong><hr></blockquote>
Actually, I was under the assumption that Intel was hearing a lot of critisim that the Plll was actaully a better chip at lower speeds then the Pllll and that Intel's next P5 or whatever was going to come in at much lower speeds. If so, then the mhz myth may not be an issue unless AMD tries to play off it.
<strong>
Uhmmm, actually they do. IBM's research labs have announced several advances that are capable of making each transistor "do more" -- or more accurately, do the same with less power.
</strong><hr></blockquote>
Not the same thing... they will still require high feature counts to implement functionality and that is going to result in big chips (and high power consumption at the targeted 2 GHz speed). Yes it will be efficient for a huge 2 GHz chip, but it will still be aimed at desktops and low-end servers.
[quote]<strong>
Nonsense. The ISA is why the P4 is as complex as it is, not the capabiliy of the chip. Intel engineers made several tradeoffs for clock speed, power consumption being just one of them. There is no reason to suppose that IBM had to make the same compromises in their design. For all you know, this chip runs cooler or as cool as the MPC7455. (OK, probably not, but the point is that we don't know and won't know until IBM releases it's thermal profile.)
</strong><hr></blockquote>
Actually from what I've read & been told the x86 decoding isn't really that much of the chip. Consider that the P-II, P-III, K7 have all done x86 instruction decoding quite effectively on much smaller transistor budgets. The cost of this in the PentiumIV didn't increase -- instead Intel threw more transistors at longer pipelines, bigger caches, more internal tables and registers, etc. As chips get bigger the fraction of the chip required to decode the x86 instructions goes down. And if you look at the POWER4 design you'll see that IBM is decoding (or cracking) their ISA in a similar manner, with similar associated costs in terms of circuitry.
I'd guess that IBM is aiming somewhere in the 60-100 million transistor range, and even on the better and considerably smaller process they are likely to use this will result in a hotter chip than the 7455... and we've yet to see Moto produce a 0.13 (or smaller) 74xx which will run much cooler if not faster.
The original statement I objected to said that Apple should move their entire lineup to the new IBM chip, if only to symbolically dump Motorola. This is misguided -- Motorola decided that aiming their top PowerPC at the embedded market was a better business move than aiming it at the top of the desktop market. Given that they have only one desktop customer, and that that customer is less than 5% of the desktop market this makes a good deal of sense, even though it is unfortunate for Apple's desktop machines. With a market of that size it just isn't worth Motorola's investment to try and compete with Intel and AMD. The embedded purposed designs are good for portables and other low-power requirements though.
IBM, on the other hand, has its own uses for desktop chips and Apple therefore becomes a way to dramatically increase the size of the market for their chips. They have already sunk a huge investment into POWER4 for their very high end machines so now it only makes sense to leverage that investment into other markets in order to recoup the original R&D cost. Now that Apple has a solid Unix-based OS I also think that their is an opportunity for IBM to broaden the appeal of their machines beyond the AIX & Linux markets, assuming that they can convince Apple that it is also in Apple's best interests (which it is). Motorola just isn't the right partner for Apple in their high end -- IBM is. Having multiple partners was the whole reason for AIM in the first place, so it is good to see that they appear to be getting back on track. Things are looking a lot more promising than they have in the last couple of years.
In term of chip performance we have been disapointed this last couple of years by the chips used in Mac.
<strong>The original statement I objected to said that Apple should move their entire lineup to the new IBM chip, if only to symbolically dump Motorola. This is misguided</strong><hr></blockquote>
I see. And yes, you're correct of course. I'd be insane to make product decisions on any other basis than what makes the most business sense.
[quote]<strong>With a market of that size it just isn't worth Motorola's investment to try and compete with Intel and AMD. The embedded purposed designs are good for portables and other low-power requirements though.</strong><hr></blockquote>
Again I agree although I would just caution that the PPC line does a good job of blurring the distinction between desktop and laptop when it comes to thermal characteristics so thinking in terms of "embedded vs desktop" may put you into a box that doesn't really exist. IBM is very power conscious in their designs and they sell to the embedded market also so they may well have had an eye to the high end of the embedded market as they scaled down the Power4 design. This chip could be in laptops 6 mos after the first version ships. (Maybe.)
[quote]<strong>Things are looking a lot more promising than they have in the last couple of years.</strong><hr></blockquote>
Absolutely. Moto is not going away and IBM is coming out with new products for the Mac. I haven't been this encouraged since the PPC first beat the P5. As someone once said, between IBM and Motorola the roadmap looks very interesting.
As for heat and power consumption, fsck all that! Who gives a damn if an iMac uses an extra 30 watts? It's NOTHING compared to the energy wasted by driving SUVs...and most people seem to LOVE making up for small dicks with giant SUVs. There's no reason Apple couldn't use quality fans in the iMac and laptops so that they don't sound like jet engines. In fact the current iMac's fan is barely audible. Drop a 2 GHz Power 5 in that iMac and then MAYBE we might actually be able to hear the fan! Oh my!
<strong>As for heat and power consumption, fsck all that! Who gives a damn if an iMac uses an extra 30 watts?</strong><hr></blockquote>
Apple's industrial designers do. An extra 30 watts from the CPU would mean three times more heat to dissipate, which makes it much harder to stuff it into tight spaces like the iMac, the eMac and the PowerBook.
It'll trickle down to those machines after a die shrink or two. In the meantime, the G4 seems to be scaling nicely. At last. A move to an onboard memory controller and RapidIO would give it a new lease on life.
<strong> Maybe some Mac users would whine about their Macs being too fast?
</strong><hr></blockquote>
"... hello, tech-support, my shuffle-puck is too fast"
It can happen, it probably already has.
<strong>
In the meantime, the G4 seems to be scaling nicely. At last.</strong><hr></blockquote>
It does? In February (actually 1/28) they shipped 1 Gig machines. Six months later they announced 1.25 Gig machines but I have yet to see one. Increasing 25% every six months is only a little worse than Moore's Law, but they don't even seem to be able to do that.
<strong>
"... hello, tech-support, my shuffle-puck is too fast"
It can happen, it probably already has.</strong><hr></blockquote>
<img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> <img src="graemlins/lol.gif" border="0" alt="[Laughing]" />
<strong>Again I agree although I would just caution that the PPC line does a good job of blurring the distinction between desktop and laptop when it comes to thermal characteristics so thinking in terms of "embedded vs desktop" may put you into a box that doesn't really exist. IBM is very power conscious in their designs and they sell to the embedded market also so they may well have had an eye to the high end of the embedded market as they scaled down the Power4 design. This chip could be in laptops 6 mos after the first version ships. (Maybe.)
</strong><hr></blockquote>
We haven't seen what a no-holds-barred desktop PowerPC can do since the 604e & 620. The best situation would be for IBM to target the limit of what next year's process can do, allowing Apple to build the fastest possible PowerPC at the time. It will take (slightly) longer to bring it down the line, but it jumps as far ahead as possible and gives them more time for the next iteration.
[ 09-04-2002: Message edited by: Programmer ]</p>
<strong>
We haven't seen what a no-holds-barred desktop PowerPC can do since the 604e & 620. The best situation would be for IBM to target the limit of what next year's process can do, allowing Apple to build the fastest possible PowerPC at the time. It will take (slightly) longer to bring it down the line, but it jumps as far ahead as possible and gives them more time for the next iteration.
[ 09-04-2002: Message edited by: Programmer ]</strong><hr></blockquote>
ha! this should have been done last year! so how long to wait? again a year ? or two? ... ok - IF it's right an Mot will use .09 and RapidIO we have the time we need. but with the current G4's at .18 and still NOT at 1.25Ghz... we don't have the time. apple can't stay at 2*1Ghz for 7 months and at 2*1.25 Ghz another 7 months... if that happens, i'd be very dissapointed.
see: i'm a freak - i wanted to buy a new computer at the time it delivers 3x performance in every area of my previous machine. with the change from the 9500/160Mhz 604e and 96MB Ram to my G4/400 i did a good jump... but now. 100Mhz bus to 166Mhz bus speed. ATA66 to ATA100 etc... i don't see that it's worth it... so. i've to wait at least another 7-8 months to buy a new killer-machine (and remember: i don't buy the fastest and most expensive machine but the one with the best price/performance (mostly the cheapest PowerMac)
[ 09-05-2002: Message edited by: Krassy ]</p>
<strong>
We haven't seen what a no-holds-barred desktop PowerPC can do since the 604e & 620. The best situation would be for IBM to target the limit of what next year's process can do, allowing Apple to build the fastest possible PowerPC at the time. It will take (slightly) longer to bring it down the line, but it jumps as far ahead as possible and gives them more time for the next iteration.
[ 09-04-2002: Message edited by: Programmer ]</strong><hr></blockquote>
We havent seen an equivelent to the 604e becouse Apple killed the clones, and as a result IBM stoped development of the CHRP. Now that Linex is maturing, and IBM is investing lots of $ in it, we might see the rebirth of the CHRP, and as a result more computers bieng sold that use PPC chips. This could only bring more revinue for IBM and Motorolla as the PPC becomes a more viable competator to the Intel/AMD offerings. Apple will be able to take advantage of it, but I doubt it will be due to number of computers that Apple sells, or for that matter Apple and IBM combined.