Bah, Nr9 was a knucklehead then and still is. 440 designs were, as Programmer has stated, sold off. Yep, quad 440 cores just ain't gonna see the light of day. That was then (and a load of crap back then) this is now.
I agree. However, some points, especially brought out by Amorph seem close to what the Cell is described as.
Quote:
Nr9
"it is a 32 bit processor it will ship next year"
Cell is 64 bit, big miss here. Nr9 also missed on the speed, cache and numerous other points.
Quote:
Nr9
"There are 4 cores so it not slow"
Not 4 cores but one PPE and 8 SPE's.
Quote:
Nr9
"it has a 440 core, a 440 FPU core, and an altivec core as one processor"
"it has an auxiliary core port
this auxiliary core is bigger than the current 440 FPU and it has altivec"
Nr9 backtracked here, but ended up saying a bigger core. Cell doesn't have what he calls an auxillary core, but the PPE is bigger than the current 440 and does have altivec.
Quote:
Nr9
"that is only 1 core though, no FPU and altivec.
however it is 1 Watt at 0.15 micron
but its also slow at 400Mhz"
SPE's have no FPU,??, but are SIMD(although not Altivec). Speed is way off though.
Quote:
Amorph
"Now, we're talking about 4 cores, 4 AltiVec units, previously unheard-of bandwidth and latency between CPUs (on MCMs) - all requiring maybe 8 watts? They can easily grow into a 64-bit variant, since the PowerPC spec is natively 64 bit."
Unheard of bandwidth fits the Cell.
I'm not saying that Nr9 knew what was going to appear "next year"(re: 2004), because obviously it didn't. But he may have heard some 2nd/3rd/4th hand information about prototypes of Cell-like cpu's under evaluation. And quite a bit of the information was garbled.
Anyway, it is interesting that the same software issues that many were debating, the suitability of this architecture in general purpose or desktop use, back then is resurfacing now. However, now we have both Sony and IBM claiming workstations based on Cell will be built.
I'm not saying that Nr9 knew what was going to appear "next year"(re: 2004), because obviously it didn't. But he may have heard some 2nd/3rd/4th hand information about prototypes of Cell-like cpu's under evaluation. And quite a bit of the information was garbled.
I wouldn't bother putting credence in anything he said. The only thing remotely interesting was the clock rate discussion, and in that he was far to extreme.
Quote:
Anyway, it is interesting that the same software issues that many were debating, the suitability of this architecture in general purpose or desktop use, back then is resurfacing now. However, now we have both Sony and IBM claiming workstations based on Cell will be built. [/B]
The reason these things "resurface" is because they are naturally recurring themes. The direction all microprocessors are going is multi-core (even the ultimate uber-core: IA-64), and you can only benefit so much from a purely symmetrical arrangement on one chip, ergo you build a chip design with specialized cores. The next questions are immediately what does the software look like and where can I use it?
If you want to talk about interesting things written in the past, go read IBM research papers published over the last 10 years on this general subject. There were plenty of clues that lead directly here, and some things which point to the future.
Yes but what kind of workstation are we talking about? It seems to me that you should be able to kill on scientific computing, and calculations on one of these like no other, but that could still make it a very limited machine in short term. I don't think a full strength GUI OS like Mac OS, or Windows would run under it. Not to mention all the apps that go with them.
You are probably correct with the first gen Cell. But how hard can it be for IBM to modify the PPE to adapt it to the platform desired? Won't there eventually be several flavors?
. . . Look at what it and its SPEs do with properly optimized vector code, however, and it'll beat pretty much anything around today (or on the x86 / PPC roadmaps)... by a significant margin.
Thank you for clearing up many issues about the cell, in this and your other posts too. It sounds like simplification of the main PPC core was a good trade-off, based on realities of the smaller 90 and 65 nm processes. The one issue not covered, or I may have missed it, is what about the FPU? From what I have read so far, my best guess is that a Cell chip does not have one, as we know it from the G5. Floating point operations are done by the SPE cores. Is this true?
This raises the software issue. How much is involved in going with a Cell based Mac? It sounds like floating point operations in existing applications would be handles like they are on a G3. Floating point would be slow. Once applications are updated, however, floating point operations should fly.
If all these things are true, I see no reason whatsoever for not going with the cell as soon as possible in all Macs. No?
You are probably correct with the first gen Cell. But how hard can it be for IBM to modify the PPE to adapt it to the platform desired? Won't there eventually be several flavors?
I have no doubts that within the next few years cell processors will be what everyone is talking about. But this prototype has shown at the least proof of concept, which is exactly what I think its supposed to do. Although I don't think this particular design (the prototype) will be the exact design to go in the playstation3, or anything else. Unless they took it as far as the playstaton prototype processor already, but I did not read that, (or forgot it) and I would have thought they would have said it. (stressed it actually)
I think this design is a middle of the road design with a fork, designed to show two things (2 mass markets potentials), Actually 3 including the playstation, but that goes without saying. By that (middle of the road with a fork) I mean that it's showing this design can potentially be used for personal computing if taken further down that path, and it is also a stellar processor for future appliances. By that I'm not saying, or referring to your refrigerator freezer, washer, and dryer but it could. And will probably be the processor of choice for those appliances. One processor with multiple cores could control the majority of your future home interacting with an OS in your refrigerator, another in your washer, and dryer, as you have probably seen on tv, But without the need for all these appliances to be made by the same manufacturer to communicate with one another, or running under a ".net", or GE, or some other Microsoft "wanna rule the world, and everything you have" control system because it can "support multiple operating systems and programming models through the use of virtualization technologies"
No doubt this is in many ways part of the future, and a new digital lifestyle.
But it's not immediately going to be in any Mac's right away.
Although with these appliances, and the cell processor design as a centerpiece, any home PC network under any OS using any type of processor be it IBM, intel, AND Texas instruments, (anything) should be able to communicate with your home through your car, phone, computer, data pad, whatever without the need to have Microsoft, Bill Gates/Josef Stalin, or some other control freak who wants to be in charge of your entire life to use it.
. . . this prototype has shown at the least proof of concept, which is exactly what I think its supposed to do.
Contraire. This chip will be manufactured in several months. A few minor tweaks to the design is the most that would happen now. The big step will be getting it working on the 65 nm process.
OK, finally made time to pull up the story I referred to from CNet..here is a key portion:
Quote:
Still, whether the chip will be able to enter different markets is another question that hinges on factors such as:
Size: Cell contains 234 million transistors and takes up 221 square millimeters in the 90-nanometer production process. That's about double the size of the 90-nanometer 3.6GHz Pentium 4, with 112 square millimeters and 125 million transistors.
Why invite your rival to your top-secret design meetings?
Big chips cost more to produce, can hide more bugs and can be tough to cram into portable devices. Cell will get cheaper when it goes to 65-nanometer production, but so will the alternatives.
Cost: Remember liquid crystal on silicon (LCOS)? The chip that would bring down the price of big-screen TVs? Intel and Brilliant Technologies failed at it. JVC and Sony succeeded. However, the latter two companies sell their LCOS chips to their own television units. The cost of the chip gets absorbed into the TV set.
Sony, Toshiba and IBM don't have to worry about the cost of Cell because they will sell it to themselves. It becomes part of a product that is tagged at a slightly higher price. An expensive Cell, however, will be a tough sell to any other manufacturers.
Alliances: Consumer electronics companies won't want to buy a processor from Sony and Toshiba. Similarly, not a lot of server manufacturers will line up to buy a Cell server chip from IBM. Why invite your rival to your top-secret design meetings?
Power: Cell will have to be air-cooled, IBM said. In other words, fans will probably be required. Ever talk on a cell phone with a fan?
While IBM didn't disclose the exact heat statistics, some at ISSCC said it could run as hot as 130 watts, more than most desktop and notebook chips. If Cell is in this range, kids will really be huddled around the PlayStation 3 at Christmas--for warmth.
On the cool engineering side, however, the chip will come with 10 digital heat sensors to warn of problems and another sensor to regulate temperature.
Memory: Cell comes with an integrated memory controller for high-performance XDR memory from Rambus--which means that the current design works exclusively with this pricey stuff. Sony used an earlier version of Rambus memory in the PlayStation, but it's been a tough sell outside of consumer electronics.
Cell is an outstanding achievement. But we have to wait and see whether it can get a job from someone other than its parents.
Heh, not really, just that Cell is still in the wait and see, laboratory stage. The press said great things about Itanium and Emotion Engine when they were announced, and where are they?
IF it does work as hoped, yes, it could be a smash hit for Apple, or whatever company uses it...
Contraire. This chip will be manufactured in several months. A few minor tweaks to the design is the most that would happen now. The big step will be getting it working on the 65 nm process.
Contraire What? The only thing contrary is what your your saying. This only further backs up what I said. do you read the whole post, or do you snip little pieces, and dismantle them as if there were no other statements furthering the view of the original statement?
"The processor shown Monday was only a prototype, and it's likely that the high-volume shipments of the processor will come when the three companies are ready to make chips using a 65-nanometer processing technology, Glaskowsky said."
"The prototype wafer and chip shown Monday were built on a 90-nanometer process technology at IBM's manufacturing facility in East Fishkill, New York. Sony will also make some of the chips at its Nagasaki, Japan, fab this year, the companies said."
Sony is going to make some of the prototype chips, but for what? They say themselves it's not going into High Volume production until it's at 65 Nanometer. Which leads me to believe it's mainly being made for further, and final R&D for reduction, and inclusion for use in the Playstation3.
You do know that an exact translation of Contraire is Opposite don't you.
It may be big stuff today, but by the time it could become useful or usable in a desktop, we should be having dual core 970 at 3-4GHz, etc.
Cell seems to be good at specific tasks, not at a general CPU like a Pentium, AMD, or PowerPC.
Maybe Apple can use them in a future MacMini or something... or maybe as a co-processor of somekind to take care of massive data crunching (rendering, dara processing, etc)
Contraire What? The only thing contrary is what your your saying. This only further backs up what I said. do you read the whole post, or do you snip little pieces, and dismantle them as if there were no other statements furthering the view of the original statement? . . .
Excuse me if I offended you. That was not my intention. I was simply disagreeing with your apparent assumption that this chip is a "proof of concept." Maybe it's partly a matter of words. In design engineering, proof of concept occurs very early in the life of a project. I believe the Cell project is at the "opposite" end of the time line, near final cleanup. (I do know the translation.) IBM and Sony are using this big and expensive Cell chips simply because nobody has 65 nm up and running reliably yet. It lets them complete the design before they have the final key component, and gives them a fast path project. There is a gamble here, betting that 65 nm will work out the way they think it will. At least that is my take on the thing. I did read your entire post but only wanted to comment on the single point that I quoted.
Re: The above quotes, I was trying to pull an interesting discussion out of a troll thread, since the thread refused to die.
What was bandied about in that thread has nothing to do with what IBM just rolled out. The PowerPC core in Cell is too different to compare directly to the current CPUs in Macs, but from 30,000 feet its performance should be comparable. Throw in the APUs and of course the whole equation changes.
Cell is not a cool CPU, and it will not get cooler going to 65nm. Apparently, the lower voltage, <4GHz variant draws 4W per APU, ie., 32 watts before you even get to the CPU or bus or cache. The higher voltage, >4GHz variant draws 11W per APU.
The particular CPU that was released is not a particularly clean or intuitive fit with Macs, mostly because of the oddball PPC. On the one hand, it could replace a G4 convincingly except for the wattage—and not coincidentally, the G4 is winding up in models where that's important. On the other hand, there is one crucial area, double-precision floating point, where the 970 looks to have the Cell PPC for lunch. A step backward in this performance metric, which is particularly important to a lot of professional and scientific customers, would not do. Cell might be able to make make up the lack of 64-bit FP capability in its PPC core with a lot of cores (4-6 Cells per machine, or variants with two PPC cores and 4 APUs?), but that's not an even tradeoff. It would have to be designed carefully and measured carefully.
That leaves: Apple putting Cell in a new class of machine (which I would not put past them at this point), or Apple putting a different variation of the Cell chip into an existing machine, or (for completeness' sake) Apple putting a different variant of Cell into a new class of machine. I like what I see of Cell, very much, but I can't fit it neatly into any of Apple's existing hardware lines. The iMac and the PowerBook would be "leading" candidates, as would the mini if Apple decides to play up the HTPC angle.
That leaves: Apple putting Cell in a new class of machine (which I would not put past them at this point), or Apple putting a different variation of the Cell chip into an existing machine, or (for completeness' sake) Apple putting a different variant of Cell into a new class of machine. I like what I see of Cell, very much, but I can't fit it neatly into any of Apple's existing hardware lines. The iMac and the PowerBook would be "leading" candidates, as would the mini if Apple decides to play up the HTPC angle.
I could see Apple using this chip to compliment the 9XX processors. I could see Apple using this chip in a video board. I could also see Apple using something like this for scientific or heavy math work. I could also see Apple being interested in this from a design concept in that Apple may want IBM to make a similar chip using VMX units. The problem would be feeding the beast. A similar chip with 4 VMX cores would a great help for code that relies heavily on vector calcs. I guess I see this chip or something like it as a great support chip, while staying with the 9XX series because the cost of moving the code base to Cell. I see Apple easing into that water not jumping in. Maybe a Cell processor will be in an Apple machine soon 2006, but in the role of video accelerator, or math/heavy calculation support if the market is there.
Heh, not really, just that Cell is still in the wait and see, laboratory stage. The press said great things about Itanium and Emotion Engine when they were announced, and where are they?
Where is the Emotion Engine?
The Emotion Engine went on to power 80+ million PS2s and has since grown up and now goes by the name Cell...
The Emotion Engine went on to power 80+ million PS2s and has since grown up and now goes by the name Cell...
True, I have one of those 80 million sitting under my TV; I meant it never made it into any other applications, and cell is not likely to either, for reasons already mentioned in this thread. (see Amorph's post)
True, I have one of those 80 million sitting under my TV; I meant it never made it into any other applications, and cell is not likely to either, for reasons already mentioned in this thread. (see Animorph's post)
Sony, IBM, and Toshiba have wide-ranging plans for Cell, from the smallest to largest scale of computing products.
Comments
Originally posted by Programmer
This PPC440 speculation is just nonsense...
I agree. However, some points, especially brought out by Amorph seem close to what the Cell is described as.
Nr9
"it is a 32 bit processor it will ship next year"
Cell is 64 bit, big miss here. Nr9 also missed on the speed, cache and numerous other points.
Nr9
"There are 4 cores so it not slow"
Not 4 cores but one PPE and 8 SPE's.
Nr9
"it has a 440 core, a 440 FPU core, and an altivec core as one processor"
"it has an auxiliary core port
this auxiliary core is bigger than the current 440 FPU and it has altivec"
Nr9 backtracked here, but ended up saying a bigger core. Cell doesn't have what he calls an auxillary core, but the PPE is bigger than the current 440 and does have altivec.
Nr9
"that is only 1 core though, no FPU and altivec.
however it is 1 Watt at 0.15 micron
but its also slow at 400Mhz"
SPE's have no FPU,??, but are SIMD(although not Altivec). Speed is way off though.
Amorph
"Now, we're talking about 4 cores, 4 AltiVec units, previously unheard-of bandwidth and latency between CPUs (on MCMs) - all requiring maybe 8 watts? They can easily grow into a 64-bit variant, since the PowerPC spec is natively 64 bit."
Unheard of bandwidth fits the Cell.
I'm not saying that Nr9 knew what was going to appear "next year"(re: 2004), because obviously it didn't. But he may have heard some 2nd/3rd/4th hand information about prototypes of Cell-like cpu's under evaluation. And quite a bit of the information was garbled.
Anyway, it is interesting that the same software issues that many were debating, the suitability of this architecture in general purpose or desktop use, back then is resurfacing now. However, now we have both Sony and IBM claiming workstations based on Cell will be built.
Originally posted by rickag
I'm not saying that Nr9 knew what was going to appear "next year"(re: 2004), because obviously it didn't. But he may have heard some 2nd/3rd/4th hand information about prototypes of Cell-like cpu's under evaluation. And quite a bit of the information was garbled.
I wouldn't bother putting credence in anything he said. The only thing remotely interesting was the clock rate discussion, and in that he was far to extreme.
Anyway, it is interesting that the same software issues that many were debating, the suitability of this architecture in general purpose or desktop use, back then is resurfacing now. However, now we have both Sony and IBM claiming workstations based on Cell will be built. [/B]
The reason these things "resurface" is because they are naturally recurring themes. The direction all microprocessors are going is multi-core (even the ultimate uber-core: IA-64), and you can only benefit so much from a purely symmetrical arrangement on one chip, ergo you build a chip design with specialized cores. The next questions are immediately what does the software look like and where can I use it?
If you want to talk about interesting things written in the past, go read IBM research papers published over the last 10 years on this general subject. There were plenty of clues that lead directly here, and some things which point to the future.
Originally posted by onlooker
Yes but what kind of workstation are we talking about? It seems to me that you should be able to kill on scientific computing, and calculations on one of these like no other, but that could still make it a very limited machine in short term. I don't think a full strength GUI OS like Mac OS, or Windows would run under it. Not to mention all the apps that go with them.
You are probably correct with the first gen Cell. But how hard can it be for IBM to modify the PPE to adapt it to the platform desired? Won't there eventually be several flavors?
Originally posted by Programmer
. . . Look at what it and its SPEs do with properly optimized vector code, however, and it'll beat pretty much anything around today (or on the x86 / PPC roadmaps)... by a significant margin.
Thank you for clearing up many issues about the cell, in this and your other posts too. It sounds like simplification of the main PPC core was a good trade-off, based on realities of the smaller 90 and 65 nm processes. The one issue not covered, or I may have missed it, is what about the FPU? From what I have read so far, my best guess is that a Cell chip does not have one, as we know it from the G5. Floating point operations are done by the SPE cores. Is this true?
This raises the software issue. How much is involved in going with a Cell based Mac? It sounds like floating point operations in existing applications would be handles like they are on a G3. Floating point would be slow. Once applications are updated, however, floating point operations should fly.
If all these things are true, I see no reason whatsoever for not going with the cell as soon as possible in all Macs. No?
Originally posted by Dave J
You are probably correct with the first gen Cell. But how hard can it be for IBM to modify the PPE to adapt it to the platform desired? Won't there eventually be several flavors?
I have no doubts that within the next few years cell processors will be what everyone is talking about. But this prototype has shown at the least proof of concept, which is exactly what I think its supposed to do. Although I don't think this particular design (the prototype) will be the exact design to go in the playstation3, or anything else. Unless they took it as far as the playstaton prototype processor already, but I did not read that, (or forgot it) and I would have thought they would have said it. (stressed it actually)
I think this design is a middle of the road design with a fork, designed to show two things (2 mass markets potentials), Actually 3 including the playstation, but that goes without saying. By that (middle of the road with a fork) I mean that it's showing this design can potentially be used for personal computing if taken further down that path, and it is also a stellar processor for future appliances. By that I'm not saying, or referring to your refrigerator freezer, washer, and dryer but it could. And will probably be the processor of choice for those appliances. One processor with multiple cores could control the majority of your future home interacting with an OS in your refrigerator, another in your washer, and dryer, as you have probably seen on tv, But without the need for all these appliances to be made by the same manufacturer to communicate with one another, or running under a ".net", or GE, or some other Microsoft "wanna rule the world, and everything you have" control system because it can "support multiple operating systems and programming models through the use of virtualization technologies"
No doubt this is in many ways part of the future, and a new digital lifestyle.
But it's not immediately going to be in any Mac's right away.
Although with these appliances, and the cell processor design as a centerpiece, any home PC network under any OS using any type of processor be it IBM, intel, AND Texas instruments, (anything) should be able to communicate with your home through your car, phone, computer, data pad, whatever without the need to have Microsoft, Bill Gates/Josef Stalin, or some other control freak who wants to be in charge of your entire life to use it.
Originally posted by onlooker
. . . this prototype has shown at the least proof of concept, which is exactly what I think its supposed to do.
Contraire. This chip will be manufactured in several months. A few minor tweaks to the design is the most that would happen now. The big step will be getting it working on the 65 nm process.
Still, whether the chip will be able to enter different markets is another question that hinges on factors such as:
Size: Cell contains 234 million transistors and takes up 221 square millimeters in the 90-nanometer production process. That's about double the size of the 90-nanometer 3.6GHz Pentium 4, with 112 square millimeters and 125 million transistors.
Why invite your rival to your top-secret design meetings?
Big chips cost more to produce, can hide more bugs and can be tough to cram into portable devices. Cell will get cheaper when it goes to 65-nanometer production, but so will the alternatives.
Cost: Remember liquid crystal on silicon (LCOS)? The chip that would bring down the price of big-screen TVs? Intel and Brilliant Technologies failed at it. JVC and Sony succeeded. However, the latter two companies sell their LCOS chips to their own television units. The cost of the chip gets absorbed into the TV set.
Sony, Toshiba and IBM don't have to worry about the cost of Cell because they will sell it to themselves. It becomes part of a product that is tagged at a slightly higher price. An expensive Cell, however, will be a tough sell to any other manufacturers.
Alliances: Consumer electronics companies won't want to buy a processor from Sony and Toshiba. Similarly, not a lot of server manufacturers will line up to buy a Cell server chip from IBM. Why invite your rival to your top-secret design meetings?
Power: Cell will have to be air-cooled, IBM said. In other words, fans will probably be required. Ever talk on a cell phone with a fan?
While IBM didn't disclose the exact heat statistics, some at ISSCC said it could run as hot as 130 watts, more than most desktop and notebook chips. If Cell is in this range, kids will really be huddled around the PlayStation 3 at Christmas--for warmth.
On the cool engineering side, however, the chip will come with 10 digital heat sensors to warn of problems and another sensor to regulate temperature.
Memory: Cell comes with an integrated memory controller for high-performance XDR memory from Rambus--which means that the current design works exclusively with this pricey stuff. Sony used an earlier version of Rambus memory in the PlayStation, but it's been a tough sell outside of consumer electronics.
Cell is an outstanding achievement. But we have to wait and see whether it can get a job from someone other than its parents.
Originally posted by iPoster
OK, finally made time to pull up the story I referred to from CNet..here is a key portion:
Sour grapes?
Originally posted by snoopy
Sour grapes?
Heh, not really, just that Cell is still in the wait and see, laboratory stage. The press said great things about Itanium and Emotion Engine when they were announced, and where are they?
IF it does work as hoped, yes, it could be a smash hit for Apple, or whatever company uses it...
Originally posted by snoopy
Contraire. This chip will be manufactured in several months. A few minor tweaks to the design is the most that would happen now. The big step will be getting it working on the 65 nm process.
Contraire What? The only thing contrary is what your your saying. This only further backs up what I said. do you read the whole post, or do you snip little pieces, and dismantle them as if there were no other statements furthering the view of the original statement?
"The processor shown Monday was only a prototype, and it's likely that the high-volume shipments of the processor will come when the three companies are ready to make chips using a 65-nanometer processing technology, Glaskowsky said."
"The prototype wafer and chip shown Monday were built on a 90-nanometer process technology at IBM's manufacturing facility in East Fishkill, New York. Sony will also make some of the chips at its Nagasaki, Japan, fab this year, the companies said."
Sony is going to make some of the prototype chips, but for what? They say themselves it's not going into High Volume production until it's at 65 Nanometer. Which leads me to believe it's mainly being made for further, and final R&D for reduction, and inclusion for use in the Playstation3.
You do know that an exact translation of Contraire is Opposite don't you.
It may be big stuff today, but by the time it could become useful or usable in a desktop, we should be having dual core 970 at 3-4GHz, etc.
Cell seems to be good at specific tasks, not at a general CPU like a Pentium, AMD, or PowerPC.
Maybe Apple can use them in a future MacMini or something... or maybe as a co-processor of somekind to take care of massive data crunching (rendering, dara processing, etc)
Originally posted by onlooker
Contraire What? The only thing contrary is what your your saying. This only further backs up what I said. do you read the whole post, or do you snip little pieces, and dismantle them as if there were no other statements furthering the view of the original statement? . . .
Excuse me if I offended you. That was not my intention. I was simply disagreeing with your apparent assumption that this chip is a "proof of concept." Maybe it's partly a matter of words. In design engineering, proof of concept occurs very early in the life of a project. I believe the Cell project is at the "opposite" end of the time line, near final cleanup. (I do know the translation.) IBM and Sony are using this big and expensive Cell chips simply because nobody has 65 nm up and running reliably yet. It lets them complete the design before they have the final key component, and gives them a fast path project. There is a gamble here, betting that 65 nm will work out the way they think it will. At least that is my take on the thing. I did read your entire post but only wanted to comment on the single point that I quoted.
Originally posted by iPoster
Heh, not really, just that Cell is still in the wait and see, laboratory stage.
Let's agree that we will disagree about what stage the project is in.
The press said great things about Itanium and Emotion Engine when they were announced, and where are they?
Yep, we shouldn't take the press seriously.
What was bandied about in that thread has nothing to do with what IBM just rolled out. The PowerPC core in Cell is too different to compare directly to the current CPUs in Macs, but from 30,000 feet its performance should be comparable. Throw in the APUs and of course the whole equation changes.
Cell is not a cool CPU, and it will not get cooler going to 65nm. Apparently, the lower voltage, <4GHz variant draws 4W per APU, ie., 32 watts before you even get to the CPU or bus or cache. The higher voltage, >4GHz variant draws 11W per APU.
The particular CPU that was released is not a particularly clean or intuitive fit with Macs, mostly because of the oddball PPC. On the one hand, it could replace a G4 convincingly except for the wattage—and not coincidentally, the G4 is winding up in models where that's important. On the other hand, there is one crucial area, double-precision floating point, where the 970 looks to have the Cell PPC for lunch. A step backward in this performance metric, which is particularly important to a lot of professional and scientific customers, would not do. Cell might be able to make make up the lack of 64-bit FP capability in its PPC core with a lot of cores (4-6 Cells per machine, or variants with two PPC cores and 4 APUs?), but that's not an even tradeoff. It would have to be designed carefully and measured carefully.
That leaves: Apple putting Cell in a new class of machine (which I would not put past them at this point), or Apple putting a different variation of the Cell chip into an existing machine, or (for completeness' sake) Apple putting a different variant of Cell into a new class of machine. I like what I see of Cell, very much, but I can't fit it neatly into any of Apple's existing hardware lines. The iMac and the PowerBook would be "leading" candidates, as would the mini if Apple decides to play up the HTPC angle.
Originally posted by Amorph
That leaves: Apple putting Cell in a new class of machine (which I would not put past them at this point), or Apple putting a different variation of the Cell chip into an existing machine, or (for completeness' sake) Apple putting a different variant of Cell into a new class of machine. I like what I see of Cell, very much, but I can't fit it neatly into any of Apple's existing hardware lines. The iMac and the PowerBook would be "leading" candidates, as would the mini if Apple decides to play up the HTPC angle.
I could see Apple using this chip to compliment the 9XX processors. I could see Apple using this chip in a video board. I could also see Apple using something like this for scientific or heavy math work. I could also see Apple being interested in this from a design concept in that Apple may want IBM to make a similar chip using VMX units. The problem would be feeding the beast. A similar chip with 4 VMX cores would a great help for code that relies heavily on vector calcs. I guess I see this chip or something like it as a great support chip, while staying with the 9XX series because the cost of moving the code base to Cell. I see Apple easing into that water not jumping in. Maybe a Cell processor will be in an Apple machine soon 2006, but in the role of video accelerator, or math/heavy calculation support if the market is there.
Originally posted by iPoster
Heh, not really, just that Cell is still in the wait and see, laboratory stage. The press said great things about Itanium and Emotion Engine when they were announced, and where are they?
Where is the Emotion Engine?
The Emotion Engine went on to power 80+ million PS2s and has since grown up and now goes by the name Cell...
Originally posted by Tuttle
Where is the Emotion Engine?
The Emotion Engine went on to power 80+ million PS2s and has since grown up and now goes by the name Cell...
True, I have one of those 80 million sitting under my TV; I meant it never made it into any other applications, and cell is not likely to either, for reasons already mentioned in this thread. (see Amorph's post)
Originally posted by iPoster
True, I have one of those 80 million sitting under my TV; I meant it never made it into any other applications, and cell is not likely to either, for reasons already mentioned in this thread. (see Animorph's post)
Sony, IBM, and Toshiba have wide-ranging plans for Cell, from the smallest to largest scale of computing products.