Don't AMD's chips do that as well? If so, why aren't they being sued?
As far as AMD is concerned they have historically reverse engineered features that they want so there may not be the necessary conditions for a lawsuit in AMDs case.
When AMD first started working on their chip designs they took a look at what Intel chips did and then reverse engineered their own work until they ended up with a work-alike.
Reverse Engineering is a common practice and generally does not cause a lawsuit. In the technology world in order for a lawsuit to work you have to show that the item in question is actually based on or derived from somebody elses work and employs demonstrable artifacts of the original.
Intel is known for having been caught with its hands in the proverbial cookie jar before. (The DEC/HP Merced/Itanium issue).
First the Cell processor would never work. Its not a general processor. It would be going back to the dark ages of CPUs. Its great for games, but not much else. As for AMD - it matters not that it was reverse engineered - that's only for copyright. Patents mean even if you figure it out yourself you aren't allowed to use it. So yes if they have patents it does not matter where it originated. If it works like it, its not allowed to be used unless licensed.
Keep in mind that Intel is simply the first target. If Transmeta won against Intel, AMD and pretty much everybody else would have to roll over and pay up on at least some of their patent portfolio.
1) You're quite right about Apple not going back to PPC but they may look at the IBM Cell chip again, I found an article talking about how the cell chip in its initial configuration was not suitable for Apple because of the power use (between 30-80 watts) but the newest version works at 25 watts and is now fully compatable with the PPC instruction set.
2) They could go to AMD. The AMD 64 bit chips run at a constant voltage regardless of the number of cores on die which makes heat management easier to design for. Apple is not a "large" purchaser of chips (unlike some other PC vendors) so it could be possible for AMD to meet demands.
Speculations R Us
Sopranino
They will NEVER look at the Cell. The greatest possibility, and is is vanishingly small since they exited the PPC, it that they might look at it for a coprocessor for some 3D operations or such. It's only ever so slightly compatable with the PPC because only the "controller" is a real PPC part. The SPE's are much simpler versions that are very limited in what they can do by themselves, and are needed if the power of the chip is to be realised. Programming the Cell is totally different from any other mainstream chip. As is the memory model.
It's a devil to program for, and I can't imagine Apple even THINKING about rewriting their OS for it. I'm not saying they couldn't do it, just that it would take a LOT of work for no good reason.
It's also a limited chip in the sense that there is no guarantee that IBM or Sony would advance the chip on a regular basis, or have more than one version out at any one time.
They would never get any serious developers to write programs for it either.
Can you honestly see MS or Adobe, or any other big developer converting their programs yet again? Especially for a chip that is so alien to every programming model they have ever used? A chip that has few tools.
No, it will never happen.
They will not be forgiven if they try to pull that.
The only reason why developers have followed them to X86 is because most developers would rather work on that chip platform since they are already using it for every other OS they are writing programs for, and they are very familiar with it, and also because there are many well developed tools and languages available.
The Cell is the unknown. No one will go there. They are having enough problems with games on the PS3 as it is.
Again, forget AMD. Perhaps in the long term they might use some AMD chips if a product requires it, and the chip has advantages. Otherwise, no. AMD simply cannot supply Apple along with Dell and their other customers.
If Intel does have this highly unlikely problem where they can't produce ships for a while, everyone will just have to wait.
In addition to not wanting to change suppliers, manufacturers will have to totally redesign their machines. Totally new mobo's will have to be designed, tested, and put into production.
You're minimizing the amount of difficulty it would take, and the cost. Apple wouldn't be the only company to attempt to get chips from AMD if they were to try to do as you say. Apple would just be another company getting in line after Dell, and hp.
As far as AMD is concerned they have historically reverse engineered features that they want so there may not be the necessary conditions for a lawsuit in AMDs case.
When AMD first started working on their chip designs they took a look at what Intel chips did and then reverse engineered their own work until they ended up with a work-alike.
Reverse Engineering is a common practice and generally does not cause a lawsuit. In the technology world in order for a lawsuit to work you have to show that the item in question is actually based on or derived from somebody elses work and employs demonstrable artifacts of the original.
Intel is known for having been caught with its hands in the proverbial cookie jar before. (The DEC/HP Merced/Itanium issue).
Sopranino
AMD has had a license from Intel for a long time. When Intel had problems in the mid '80's, IBM forced them to license out the x86 IP. Several companies manufactured x86 chips for years, including IBM itself. Most have dropped out. But AMD and Intel have numerous cross licensed bits of IP from each other. The license says that AMD can use many developments that Intel comes up with, and visa versa.
AMD has had a license from Intel for a long time. When Intel had problems in the mid '80's, IBM forced them to license out the x86 IP. Several companies manufactured x86 chips for years, including IBM itself. Most have dropped out. But AMD and Intel have numerous cross licensed bits of IP from each other. The license says that AMD can use many developments that Intel comes up with, and visa versa.
Reverse engineering isn't a question here.
Thanks for adding that information in to the mix.....you're quite right about the cross licensing part, however my original statements still stands. When AMD started out they were reverse engineering Intel designs in order to develop their own chips, this occurs prior to the events that you refer to above. What you bring to the table here is still relevant however as the question becomes whether or not the cross licensed parts of the AMD architecture have the patent infringement designs in them.
They will NEVER look at the Cell. The greatest possibility, and is is vanishingly small since they exited the PPC, it that they might look at it for a coprocessor for some 3D operations or such. It's only ever so slightly compatable with the PPC because only the "controller" is a real PPC part. The SPE's are much simpler versions that are very limited in what they can do by themselves, and are needed if the power of the chip is to be realised. Programming the Cell is totally different from any other mainstream chip. As is the memory model.
It's a devil to program for, and I can't imagine Apple even THINKING about rewriting their OS for it. I'm not saying they couldn't do it, just that it would take a LOT of work for no good reason.
According to a close aquaintance of mine who is a Tech Analyst/R&D for a major control systems manufacturer, IBM has a low level development Toolkit that does a lot of the nasty work for developers. (I'll dig up the relevant info for you)
Quote:
Originally Posted by melgross
It's also a limited chip in the sense that there is no guarantee that IBM or Sony would advance the chip on a regular basis, or have more than one version out at any one time.
IBM is already into their second generation of this chip design. There are a number of companies who are utilizing this chip (not just game consoles). IBM would not sink a large amount of money into developing a one shot production and then drop it. However in support of your statement I have not been able to find any concrete information on a long term roadmap for the cell chip.
Quote:
Originally Posted by melgross
They would never get any serious developers to write programs for it either.
Errrm, why not??
Quote:
Originally Posted by melgross
Can you honestly see MS or Adobe, or any other big developer converting their programs yet again? Especially for a chip that is so alien to every programming model they have ever used? A chip that has few tools.
This could be rectified with a translation layer added in at the beginning of the cell chip introduction. (Similar to the Rosetta layer in the current OSX environment). Here again the long term solution would ultimately ride on the developer tools, released by IBM, being able to handle the actual compiling of source code so that developers can build code as they normally would under non-cell architecture. As I pointed out above the latest developer tools from IBM seem to take care of some of this already.
Quote:
Originally Posted by melgross
No, it will never happen.
They will not be forgiven if they try to pull that.
The only reason why developers have followed them to X86 is because most developers would rather work on that chip platform since they are already using it for every other OS they are writing programs for, and they are very familiar with it, and also because there are many well developed tools and languages available.
The Cell is the unknown. No one will go there. They are having enough problems with games on the PS3 as it is.
As in anything new you will get some who resist change or will not work with it at all, and others who would dive right in. Just because something is unknown does not prevent people from going there. Yes developers are having problems with the PS3 but they are getting things done so the tool kits are being used/developed/evolved.
Quote:
Originally Posted by melgross
Again, forget AMD. Perhaps in the long term they might use some AMD chips if a product requires it, and the chip has advantages. Otherwise, no. AMD simply cannot supply Apple along with Dell and their other customers.
If Intel does have this highly unlikely problem where they can't produce ships for a while, everyone will just have to wait.
In addition to not wanting to change suppliers, manufacturers will have to totally redesign their machines. Totally new mobo's will have to be designed, tested, and put into production.
You're minimizing the amount of difficulty it would take, and the cost. Apple wouldn't be the only company to attempt to get chips from AMD if they were to try to do as you say. Apple would just be another company getting in line after Dell, and hp.
Actually I never made any statement one way or the other about difficulty. My original intent was to propose a couple of possible avenues that could be taken. I wrote the original as pure speculation and did not write an essay to try to answer all possible concerns or problems with them.
I do appreciate your observations however and I am quite impressed with the points that you have brought up.
According to a close aquaintance of mine who is a Tech Analyst/R&D for a major control systems manufacturer, IBM has a low level development Toolkit that does a lot of the nasty work for developers. (I'll dig up the relevant info for you)
Perhaps they do, but the biggest complaint is that both IBM and Sony's tools are just barely adequate. It will take years before they come even close to the rich enviornments that are available for all of the other well known chip families.
Quote:
IBM is already into their second generation of this chip design. There are a number of companies who are utilizing this chip (not just game consoles). IBM would not sink a large amount of money into developing a one shot production and then drop it. However in support of your statement I have not been able to find any concrete information on a long term roadmap for the cell chip.
There is always an assumption that a company is into a next generation of a product. When evidence of anything more than some internal documentation of a next generation wishlist come around then we can look to it, but not before.
IBM is hoping to sell perhaps 100 million of these chips over the long term into the games machine market. Those chips will be exactly the same as the chip we see now. console manufacturers do not upgrade the electronics in their generation of machines except to lower costs.
You can't find information on a roadmap, because as of now, there isn't one.
Quote:
Errrm, why not??
I explained that pretty throughly in my post.
Quote:
This could be rectified with a translation layer added in at the beginning of the cell chip introduction. (Similar to the Rosetta layer in the current OSX environment). Here again the long term solution would ultimately ride on the developer tools, released by IBM, being able to handle the actual compiling of source code so that developers can build code as they normally would under non-cell architecture. As I pointed out above the latest developer tools from IBM seem to take care of some of this already.
Translation layers don't work well. What you are talking about would work no better than Virtual PC, or the older Softwindows.
Rosetta also uses a similar scheme as you pointed out, but that doesn't work well. Not working well meaning that it is slow, and isn't command complete.
None of those methods are usable other than for a limited use product. They rob most of the performance, and can't translate all of the IP. They are NOT a viable way to allow a product such as the Cell to function at useful levels.
Quote:
As in anything new you will get some who resist change or will not work with it at all, and others who would dive right in. Just because something is unknown does not prevent people from going there. Yes developers are having problems with the PS3 but they are getting things done so the tool kits are being used/developed/evolved.
This has nothing to do with resisting change. There are practical matters that weigh in. Some of them I have answered here. Some, in my earlier post. Game machines are a fraction of the complexity of a modern computer, or its OS. There are no games that even approach the complexity of the larger programs available to computer users.
Both Vista and 10.5 approach 50 million lines of code. It's almost impossible to see how all of that could be converted to the Cell in any time scale that would be useful. The costs would be enormous. As there are almost no programmers familiar with producing code for it, the learning curve would also be incredible. There are no compilers out there right now that could assist this project in any meaningful way.
Quote:
Actually I never made any statement one way or the other about difficulty. My original intent was to propose a couple of possible avenues that could be taken. I wrote the original as pure speculation and did not write an essay to try to answer all possible concerns or problems with them.
Yes. So far, no one who has advocated this has made a good case for the practicality of it. Most make no case at all, except to say, ooh, this would be SO great. The majority of trusted people in this business have weighed in with the case of its being properly rejected by Apple.
Quote:
I do appreciate your observations however and I am quite impressed with the points that you have brought up.
Sopranino
Well thanks. I enjoy arguments that don't devolve into name calling, no matter how much we might disagree.
Perhaps they do, but the biggest complaint is that both IBM and Sony's tools are just barely adequate. It will take years before they come even close to the rich enviornments that are available for all of the other well known chip families.
There is always an assumption that a company is into a next generation of a product. When evidence of anything more than some internal documentation of a next generation wishlist come around then we can look to it, but not before.
IBM is hoping to sell perhaps 100 million of these chips over the long term into the games machine market. Those chips will be exactly the same as the chip we see now. console manufacturers do not upgrade the electronics in their generation of machines except to lower costs.
You can't find information on a roadmap, because as of now, there isn't one.
I explained that pretty throughly in my post.
Translation layers don't work well. What you are talking about would work no better than Virtual PC, or the older Softwindows.
Rosetta also uses a similar scheme as you pointed out, but that doesn't work well. Not working well meaning that it is slow, and isn't command complete.
None of those methods are usable other than for a limited use product. They rob most of the performance, and can't translate all of the IP. They are NOT a viable way to allow a product such as the Cell to function at useful levels.
This has nothing to do with resisting change. There are practical matters that weigh in. Some of them I have answered here. Some, in my earlier post. Game machines are a fraction of the complexity of a modern computer, or its OS. There are no games that even approach the complexity of the larger programs available to computer users.
Both Vista and 10.5 approach 50 million lines of code. It's almost impossible to see how all of that could be converted to the Cell in any time scale that would be useful. The costs would be enormous. As there are almost no programmers familiar with producing code for it, the learning curve would also be incredible. There are no compilers out there right now that could assist this project in any meaningful way.
Yes. So far, no one who has advocated this has made a good case for the practicality of it. Most make no case at all, except to say, ooh, this would be SO great. The majority of trusted people in this business have weighed in with the case of its being properly rejected by Apple.
Well thanks. I enjoy arguments that don't devolve into name calling, no matter how much we might disagree.
You bring up some very nicely reasoned comments, although I do agree with some of them there are some other ones that I think need some clarification (from my side if nothing else). I'll post my responses to those that I have questions about a little later. (I'm really tired, been up all day prepping for a recital) Like yourself I enjoy a good discussion no matter if the points of view differ or not. The purpose behind these posts is to engender intelligent speculation. Thankyou......
You bring up some very nicely reasoned comments, although I do agree with some of them there are some other ones that I think need some clarification (from my side if nothing else). I'll post my responses to those that I have questions about a little later. (I'm really tired, been up all day prepping for a recital) Like yourself I enjoy a good discussion no matter if the points of view differ or not. The purpose behind these posts is to engender intelligent speculation. Thankyou......
Sopranino (tired now, good night)
Lights out!
You might want to read this article. It gives some perspective on compilers.
Comments
Don't AMD's chips do that as well? If so, why aren't they being sued?
As far as AMD is concerned they have historically reverse engineered features that they want so there may not be the necessary conditions for a lawsuit in AMDs case.
When AMD first started working on their chip designs they took a look at what Intel chips did and then reverse engineered their own work until they ended up with a work-alike.
Reverse Engineering is a common practice and generally does not cause a lawsuit. In the technology world in order for a lawsuit to work you have to show that the item in question is actually based on or derived from somebody elses work and employs demonstrable artifacts of the original.
Intel is known for having been caught with its hands in the proverbial cookie jar before. (The DEC/HP Merced/Itanium issue).
Sopranino
I guess several options could exist.
1) You're quite right about Apple not going back to PPC but they may look at the IBM Cell chip again, I found an article talking about how the cell chip in its initial configuration was not suitable for Apple because of the power use (between 30-80 watts) but the newest version works at 25 watts and is now fully compatable with the PPC instruction set.
2) They could go to AMD. The AMD 64 bit chips run at a constant voltage regardless of the number of cores on die which makes heat management easier to design for. Apple is not a "large" purchaser of chips (unlike some other PC vendors) so it could be possible for AMD to meet demands.
Speculations R Us
Sopranino
They will NEVER look at the Cell. The greatest possibility, and is is vanishingly small since they exited the PPC, it that they might look at it for a coprocessor for some 3D operations or such. It's only ever so slightly compatable with the PPC because only the "controller" is a real PPC part. The SPE's are much simpler versions that are very limited in what they can do by themselves, and are needed if the power of the chip is to be realised. Programming the Cell is totally different from any other mainstream chip. As is the memory model.
It's a devil to program for, and I can't imagine Apple even THINKING about rewriting their OS for it. I'm not saying they couldn't do it, just that it would take a LOT of work for no good reason.
It's also a limited chip in the sense that there is no guarantee that IBM or Sony would advance the chip on a regular basis, or have more than one version out at any one time.
They would never get any serious developers to write programs for it either.
Can you honestly see MS or Adobe, or any other big developer converting their programs yet again? Especially for a chip that is so alien to every programming model they have ever used? A chip that has few tools.
No, it will never happen.
They will not be forgiven if they try to pull that.
The only reason why developers have followed them to X86 is because most developers would rather work on that chip platform since they are already using it for every other OS they are writing programs for, and they are very familiar with it, and also because there are many well developed tools and languages available.
The Cell is the unknown. No one will go there. They are having enough problems with games on the PS3 as it is.
Again, forget AMD. Perhaps in the long term they might use some AMD chips if a product requires it, and the chip has advantages. Otherwise, no. AMD simply cannot supply Apple along with Dell and their other customers.
If Intel does have this highly unlikely problem where they can't produce ships for a while, everyone will just have to wait.
In addition to not wanting to change suppliers, manufacturers will have to totally redesign their machines. Totally new mobo's will have to be designed, tested, and put into production.
You're minimizing the amount of difficulty it would take, and the cost. Apple wouldn't be the only company to attempt to get chips from AMD if they were to try to do as you say. Apple would just be another company getting in line after Dell, and hp.
As far as AMD is concerned they have historically reverse engineered features that they want so there may not be the necessary conditions for a lawsuit in AMDs case.
When AMD first started working on their chip designs they took a look at what Intel chips did and then reverse engineered their own work until they ended up with a work-alike.
Reverse Engineering is a common practice and generally does not cause a lawsuit. In the technology world in order for a lawsuit to work you have to show that the item in question is actually based on or derived from somebody elses work and employs demonstrable artifacts of the original.
Intel is known for having been caught with its hands in the proverbial cookie jar before. (The DEC/HP Merced/Itanium issue).
Sopranino
AMD has had a license from Intel for a long time. When Intel had problems in the mid '80's, IBM forced them to license out the x86 IP. Several companies manufactured x86 chips for years, including IBM itself. Most have dropped out. But AMD and Intel have numerous cross licensed bits of IP from each other. The license says that AMD can use many developments that Intel comes up with, and visa versa.
Reverse engineering isn't a question here.
AMD has had a license from Intel for a long time. When Intel had problems in the mid '80's, IBM forced them to license out the x86 IP. Several companies manufactured x86 chips for years, including IBM itself. Most have dropped out. But AMD and Intel have numerous cross licensed bits of IP from each other. The license says that AMD can use many developments that Intel comes up with, and visa versa.
Reverse engineering isn't a question here.
Thanks for adding that information in to the mix.....you're quite right about the cross licensing part, however my original statements still stands. When AMD started out they were reverse engineering Intel designs in order to develop their own chips, this occurs prior to the events that you refer to above. What you bring to the table here is still relevant however as the question becomes whether or not the cross licensed parts of the AMD architecture have the patent infringement designs in them.
Sopranino
They will NEVER look at the Cell. The greatest possibility, and is is vanishingly small since they exited the PPC, it that they might look at it for a coprocessor for some 3D operations or such. It's only ever so slightly compatable with the PPC because only the "controller" is a real PPC part. The SPE's are much simpler versions that are very limited in what they can do by themselves, and are needed if the power of the chip is to be realised. Programming the Cell is totally different from any other mainstream chip. As is the memory model.
It's a devil to program for, and I can't imagine Apple even THINKING about rewriting their OS for it. I'm not saying they couldn't do it, just that it would take a LOT of work for no good reason.
According to a close aquaintance of mine who is a Tech Analyst/R&D for a major control systems manufacturer, IBM has a low level development Toolkit that does a lot of the nasty work for developers. (I'll dig up the relevant info for you)
It's also a limited chip in the sense that there is no guarantee that IBM or Sony would advance the chip on a regular basis, or have more than one version out at any one time.
IBM is already into their second generation of this chip design. There are a number of companies who are utilizing this chip (not just game consoles). IBM would not sink a large amount of money into developing a one shot production and then drop it. However in support of your statement I have not been able to find any concrete information on a long term roadmap for the cell chip.
They would never get any serious developers to write programs for it either.
Errrm, why not??
Can you honestly see MS or Adobe, or any other big developer converting their programs yet again? Especially for a chip that is so alien to every programming model they have ever used? A chip that has few tools.
This could be rectified with a translation layer added in at the beginning of the cell chip introduction. (Similar to the Rosetta layer in the current OSX environment). Here again the long term solution would ultimately ride on the developer tools, released by IBM, being able to handle the actual compiling of source code so that developers can build code as they normally would under non-cell architecture. As I pointed out above the latest developer tools from IBM seem to take care of some of this already.
No, it will never happen.
They will not be forgiven if they try to pull that.
The only reason why developers have followed them to X86 is because most developers would rather work on that chip platform since they are already using it for every other OS they are writing programs for, and they are very familiar with it, and also because there are many well developed tools and languages available.
The Cell is the unknown. No one will go there. They are having enough problems with games on the PS3 as it is.
As in anything new you will get some who resist change or will not work with it at all, and others who would dive right in. Just because something is unknown does not prevent people from going there. Yes developers are having problems with the PS3 but they are getting things done so the tool kits are being used/developed/evolved.
Again, forget AMD. Perhaps in the long term they might use some AMD chips if a product requires it, and the chip has advantages. Otherwise, no. AMD simply cannot supply Apple along with Dell and their other customers.
If Intel does have this highly unlikely problem where they can't produce ships for a while, everyone will just have to wait.
In addition to not wanting to change suppliers, manufacturers will have to totally redesign their machines. Totally new mobo's will have to be designed, tested, and put into production.
You're minimizing the amount of difficulty it would take, and the cost. Apple wouldn't be the only company to attempt to get chips from AMD if they were to try to do as you say. Apple would just be another company getting in line after Dell, and hp.
Actually I never made any statement one way or the other about difficulty. My original intent was to propose a couple of possible avenues that could be taken. I wrote the original as pure speculation and did not write an essay to try to answer all possible concerns or problems with them.
I do appreciate your observations however and I am quite impressed with the points that you have brought up.
Sopranino
According to a close aquaintance of mine who is a Tech Analyst/R&D for a major control systems manufacturer, IBM has a low level development Toolkit that does a lot of the nasty work for developers. (I'll dig up the relevant info for you)
Perhaps they do, but the biggest complaint is that both IBM and Sony's tools are just barely adequate. It will take years before they come even close to the rich enviornments that are available for all of the other well known chip families.
IBM is already into their second generation of this chip design. There are a number of companies who are utilizing this chip (not just game consoles). IBM would not sink a large amount of money into developing a one shot production and then drop it. However in support of your statement I have not been able to find any concrete information on a long term roadmap for the cell chip.
There is always an assumption that a company is into a next generation of a product. When evidence of anything more than some internal documentation of a next generation wishlist come around then we can look to it, but not before.
IBM is hoping to sell perhaps 100 million of these chips over the long term into the games machine market. Those chips will be exactly the same as the chip we see now. console manufacturers do not upgrade the electronics in their generation of machines except to lower costs.
You can't find information on a roadmap, because as of now, there isn't one.
Errrm, why not??
I explained that pretty throughly in my post.
This could be rectified with a translation layer added in at the beginning of the cell chip introduction. (Similar to the Rosetta layer in the current OSX environment). Here again the long term solution would ultimately ride on the developer tools, released by IBM, being able to handle the actual compiling of source code so that developers can build code as they normally would under non-cell architecture. As I pointed out above the latest developer tools from IBM seem to take care of some of this already.
Translation layers don't work well. What you are talking about would work no better than Virtual PC, or the older Softwindows.
Rosetta also uses a similar scheme as you pointed out, but that doesn't work well. Not working well meaning that it is slow, and isn't command complete.
None of those methods are usable other than for a limited use product. They rob most of the performance, and can't translate all of the IP. They are NOT a viable way to allow a product such as the Cell to function at useful levels.
As in anything new you will get some who resist change or will not work with it at all, and others who would dive right in. Just because something is unknown does not prevent people from going there. Yes developers are having problems with the PS3 but they are getting things done so the tool kits are being used/developed/evolved.
This has nothing to do with resisting change. There are practical matters that weigh in. Some of them I have answered here. Some, in my earlier post. Game machines are a fraction of the complexity of a modern computer, or its OS. There are no games that even approach the complexity of the larger programs available to computer users.
Both Vista and 10.5 approach 50 million lines of code. It's almost impossible to see how all of that could be converted to the Cell in any time scale that would be useful. The costs would be enormous. As there are almost no programmers familiar with producing code for it, the learning curve would also be incredible. There are no compilers out there right now that could assist this project in any meaningful way.
Actually I never made any statement one way or the other about difficulty. My original intent was to propose a couple of possible avenues that could be taken. I wrote the original as pure speculation and did not write an essay to try to answer all possible concerns or problems with them.
Yes. So far, no one who has advocated this has made a good case for the practicality of it. Most make no case at all, except to say, ooh, this would be SO great. The majority of trusted people in this business have weighed in with the case of its being properly rejected by Apple.
I do appreciate your observations however and I am quite impressed with the points that you have brought up.
Sopranino
Well thanks. I enjoy arguments that don't devolve into name calling, no matter how much we might disagree.
Perhaps they do, but the biggest complaint is that both IBM and Sony's tools are just barely adequate. It will take years before they come even close to the rich enviornments that are available for all of the other well known chip families.
There is always an assumption that a company is into a next generation of a product. When evidence of anything more than some internal documentation of a next generation wishlist come around then we can look to it, but not before.
IBM is hoping to sell perhaps 100 million of these chips over the long term into the games machine market. Those chips will be exactly the same as the chip we see now. console manufacturers do not upgrade the electronics in their generation of machines except to lower costs.
You can't find information on a roadmap, because as of now, there isn't one.
I explained that pretty throughly in my post.
Translation layers don't work well. What you are talking about would work no better than Virtual PC, or the older Softwindows.
Rosetta also uses a similar scheme as you pointed out, but that doesn't work well. Not working well meaning that it is slow, and isn't command complete.
None of those methods are usable other than for a limited use product. They rob most of the performance, and can't translate all of the IP. They are NOT a viable way to allow a product such as the Cell to function at useful levels.
This has nothing to do with resisting change. There are practical matters that weigh in. Some of them I have answered here. Some, in my earlier post. Game machines are a fraction of the complexity of a modern computer, or its OS. There are no games that even approach the complexity of the larger programs available to computer users.
Both Vista and 10.5 approach 50 million lines of code. It's almost impossible to see how all of that could be converted to the Cell in any time scale that would be useful. The costs would be enormous. As there are almost no programmers familiar with producing code for it, the learning curve would also be incredible. There are no compilers out there right now that could assist this project in any meaningful way.
Yes. So far, no one who has advocated this has made a good case for the practicality of it. Most make no case at all, except to say, ooh, this would be SO great. The majority of trusted people in this business have weighed in with the case of its being properly rejected by Apple.
Well thanks. I enjoy arguments that don't devolve into name calling, no matter how much we might disagree.
You bring up some very nicely reasoned comments, although I do agree with some of them there are some other ones that I think need some clarification (from my side if nothing else). I'll post my responses to those that I have questions about a little later. (I'm really tired, been up all day prepping for a recital) Like yourself I enjoy a good discussion no matter if the points of view differ or not. The purpose behind these posts is to engender intelligent speculation. Thankyou......
Sopranino (tired now, good night)
You bring up some very nicely reasoned comments, although I do agree with some of them there are some other ones that I think need some clarification (from my side if nothing else). I'll post my responses to those that I have questions about a little later. (I'm really tired, been up all day prepping for a recital) Like yourself I enjoy a good discussion no matter if the points of view differ or not. The purpose behind these posts is to engender intelligent speculation. Thankyou......
Sopranino (tired now, good night)
Lights out!
You might want to read this article. It gives some perspective on compilers.
http://www.macresearch.org/interview...compiler_group