I was looking through the newly announced 200m series cards and only the 210, 230, and 240 support OpenCL and the 230/40 both have the same power consumption as the 9600gt they succeed.
Makes you think, though the 210m isn't a motherboard, it does succeed the 9400m card, so maybe in the near future the MBPs will have a 210m motherboard and a 230/40 discrete card to replace the current setup.
The 250m gts doesn't show OpenCL compatibility on the Nvidia site, so it seems illogical. And an increase in power consumption seems silly since it would reduce the battery life claims Apple makes a big deal out of right now.
I was looking through the newly announced 200m series cards and only the 210, 230, and 240 support OpenCL and the 230/40 both have the same power consumption as the 9600gt they succeed.
Makes you think, though the 210m isn't a motherboard, it does succeed the 9400m card, so maybe in the near future the MBPs will have a 210m motherboard and a 230/40 discrete card to replace the current setup.
The 250m gts doesn't show OpenCL compatibility on the Nvidia site, so it seems illogical. And an increase in power consumption seems silly since it would reduce the battery life claims Apple makes a big deal out of right now.
Well, i am certain all G200 and G92 based Nvidia Product support OpenCL. So may be the page isn't updated yet.
The Reason why GTS 250M got my attention was its Performance / Watt. Which nearly double the currently Best Mobile Offering on the Market.
The GT240M will give 50% !!!! more performance then the current 9600M GT used in MBP at the same power usage!!!!
And GTS 250M will give MORE THEN 150% INCREASE in PEROMANCE of current 9600M GT. Using only 20% more power.
I cant wait till Apple get these update. It is a Huge Leap comparer to current offering.
I'll put my money on the GT240M then. Keeping the machines thin, cool and battery life long is essential for Apple. Also, Apple will probably reason that "less than a single digit percent of Macbook Pro users" actually need the power of a GTS250M
Don't count on it. Nvidia's next line, the GT300s, are going to be considerably more powerful than the 200 series, and are due out in the fall. Once these come out, the performance (and age) difference between the midrange 300 (probably the GT340/50M) and the current 9600M will be too large to ignore, and Apple will be forced to upgrade to it to say it has capable current generation graphics.
The good news is that for us MBP users, the difference will be like night and day.
The problem is that these hot new GPUs aren't compatible with Arrandale, with its integrated GPU and memory controller. Using them would require Apple go to a 3-chip solution from the current 2-chip one. Right now the 9400m has an integrated memory controller, so Apple only needs 2 chips (CPU + GPU), but if Apple wants to work around the crummy performance of Intel's new integrated GPU (1/2 that of the 9400m) it needs an external controller in addition to the external GPU. The extra heat and cost will not be to Apple's liking. Intel deliberately designing their CPU to shut out the competition? Say it isn't so!
The problem is that these hot new GPUs aren't compatible with Arrandale, with its integrated GPU and memory controller. Using them would require Apple go to a 3-chip solution from the current 2-chip one. Right now the 9400m has an integrated memory controller, so Apple only needs 2 chips (CPU + GPU), but if Apple wants to work around the crummy performance of Intel's new integrated GPU (1/2 that of the 9400m) it needs an external controller in addition to the external GPU. The extra heat and cost will not be to Apple's liking. Intel deliberately designing their CPU to shut out the competition? Say it isn't so!
240 and 250 will have exactly the same type of interconnection as current 9600M GT (it uses PCI-Express x16 bus and dedicated GDDR3 memory onboard).
And only difference between current and future generation will be in Nothbridge that doesn't have IMC and IGP.
Apple now uses 3 chip solution: CPU + NB (which has 9400M GT and other stuff integrated) + 9600M GT
240 and 250 will have exactly the same type of interconnection as current 9600M GT (it uses PCI-Express x16 bus and dedicated GDDR3 memory onboard).
And only difference between current and future generation will be in Nothbridge that doesn't have IMC and IGP.
Apple now uses 3 chip solution: CPU + NB (which has 9400M GT and other stuff integrated) + 9600M GT
Apple's low end laptops ( and the Mini) don't have a 9600m. Remember that Arrandale doesn't even have a Northbridge, it's all integrated on die. Since Larrabee is 45nm while Arrandale will be 32nm, my bet is the IGP resides in the same package as the CPU, but is not on die. Do you think Apple will willingly suffer 1/2 the graphics performance from Larrabee (in comparison with 9400m)? Maybe, but that will be a tough sell for its customers. That means an external IGP, but Intel claims Nvidia doesn't have the required license to integrate with the on-die IMC. While that remains in litigation, the only way Apple will be able to maintain graphics performance for products that don't have their own video memory is to use an external GPU along with an external memory controller. That's likely to be costly in heat, motherboard space, or more likely both.
Apple's low end laptops ( and the Mini) don't have a 9600m. Remember that Arrandale doesn't even have a Northbridge, it's all integrated on die. Since Larrabee is 45nm while Arrandale will be 32nm, my bet is the IGP resides in the same package as the CPU, but is not on die. Do you think Apple will willingly suffer 1/2 the graphics performance from Larrabee (in comparison with 9400m)? Maybe, but that will be a tough sell for its customers. That means an external IGP, but Intel claims Nvidia doesn't have the required license to integrate with the on-die IMC. While that remains in litigation, the only way Apple will be able to maintain graphics performance for products that don't have their own video memory is to use an external GPU along with an external memory controller. That's likely to be costly in heat, motherboard space, or more likely both.
Apple will likely use discrete G210M to complement Arrandale's IGP in low-end segment as well.
I can't imagine Apple reverting back to the poor performance of Intels chips so what is ip? AMD I don't think so. An internally built GPU, possibly.
Apple seriously needs to worry about Intels behaviour with respect to GPU producing third parties as it is not in their best interest. The risk is high that they would get locked into poor performing and buggy GPUs and be extremely limited in their system design abilities. I just see going all Intel as a serious long term mistake.
In anyevent has Nvidia been all that much of a reliability issue after this issue was identified. I'm just not hearing about excessive numbers of Macs coming back with GPU failures anymore. Considering the solid performance increases coming in the same power profiles what else is there for Apple to complain about?
As to internally produce GPUs I can see an arguement for this. Especially in the case of a GPU that is also an optimized OpenCL engine. The bad part is that it would be proprietary to Apple. The good part is that it's structure could be tightly tailored for Apples needs.
I guess time will tell here but I'd be very surprised to see Apple step away from Nvidia. The relationship seems to be firing on all cylinders right now. If it has gone bad it must be over something that is not public at the moment.
I guess time will tell here but I'd be very surprised to see Apple step away from Nvidia. The relationship seems to be firing on all cylinders right now. If it has gone bad it must be over something that is not public at the moment.
My bet is it's something very public: Nvidia's litigation against Intel. If Apple wants to get its new integrated-GPU products out on time and still use Intel's CPUs they must drop Nvidia. Litigation is so uncertain that Apple can't afford to place bets on who will win or when it will end, nor can they afford Intel refusing to ship them product, or put them on the back of the line instead of the front, because Apple continues to do business with Nvidia.
Not selling Nehalem licenses makes me assume that squeezing out third-party vendors in the integrated-GPU space was the entire point of the Arrandale design.
My bet is it's something very public: Nvidia's litigation against Intel. If Apple wants to get its new integrated-GPU products out on time and still use Intel's CPUs they must drop Nvidia. Litigation is so uncertain that Apple can't afford to place bets on who will win or when it will end, nor can they afford Intel refusing to ship them product, or put them on the back of the line instead of the front, because Apple continues to do business with Nvidia.
Not selling Nehalem licenses makes me assume that squeezing out third-party vendors in the integrated-GPU space was the entire point of the Arrandale design.
Are you sure that Arrandale will provide RAM interface for integrated videocard?
I'm pretty sure that using discrete GPU with dedicated memory is only possible solution regardless of chipset.
And I really hope Apple to use Intel chipset instead of nVidia's one combined with 210M for 13" (or AMD equivalent) and 240/250M for 15/17 (or AMD equivalent).
You'll get best of all worlds - minimum power consumption of Intel's iGP, maximum stability with Intel's chipset and still high 3D performance with powerful discrete GPU.
Seems like that guy has an issue with Nvidia himself and wants to see AMD/ATI do better. Apparently he posted something about Dell and Nvidia too, which hasn't come to anything.
The best way to get traffic to your site is to post a plausible rumor like that with no evidence that it took place and no way to prove it didn't but one that has important consequences.
The problem with the defective GPUs is quite far reaching for Apple as it could affect every single Macbook Pro shipped in the last 2-3 years but as mentioned above, there doesn't seem to be a huge return rate noted anywhere.
It's going to happen one day that Apple will drop external GPUs altogether, possibly when they hit 22nm Ivy Bridge in 2011. If Intel's chips weren't so terrible right now, I wouldn't mind. But they are, so for the time being, I'd hope any issues are dealt with.
Nvidia's developments with CUDA have been very good and show a lot of actual real-world use, unlike AMD's GPGPU developments so far. OpenCL will probably change this but that's how it is now.
One day, the aim will be that there is a single processor using fast internal Ram. This locks Nvidia out entirely unless Intel partners up with them and that looks unlikely. For AMD/ATI it doesn't matter as they do their own CPU solution anyway.
If it comes down to that level, I'd rather have an AMD/ATI solution than an Intel one with Intel graphics.
It will certainly be an interesting turn of events if lots of companies start turning their back on NVidia.
My bet is it's something very public: Nvidia's litigation against Intel. If Apple wants to get its new integrated-GPU products out on time and still use Intel's CPUs they must drop Nvidia. Litigation is so uncertain that Apple can't afford to place bets on who will win or when it will end, nor can they afford Intel refusing to ship them product, or put them on the back of the line instead of the front, because Apple continues to do business with Nvidia.
This really makes sense to me, however I still can't see Apple going back to Intel hardware for GPU's no matter how slightly improved they would be. It would be a very hard sell unless those iGPUs are a lot better than what many of us expect. That would mean good 2 & 3D support and functional video acceleration hardware. Or an architecture that can accelerate video without the need for special hardware. Intel just doesn't inspire me here.
Quote:
Not selling Nehalem licenses makes me assume that squeezing out third-party vendors in the integrated-GPU space was the entire point of the Arrandale design.
This I'm very concerned about. It is something that needs proactive regulatory action if you ask me. If nothing else it will have a dramatically negative impact on innovation and advancement of the art.
Seems like that guy has an issue with Nvidia himself and wants to see AMD/ATI do better. Apparently he posted something about Dell and Nvidia too, which hasn't come to anything.
The best way to get traffic to your site is to post a plausible rumor like that with no evidence that it took place and no way to prove it didn't but one that has important consequences.
The problem with the defective GPUs is quite far reaching for Apple as it could affect every single Macbook Pro shipped in the last 2-3 years but as mentioned above, there doesn't seem to be a huge return rate noted anywhere.
It's going to happen one day that Apple will drop external GPUs altogether, possibly when they hit 22nm Ivy Bridge in 2011. If Intel's chips weren't so terrible right now, I wouldn't mind. But they are, so for the time being, I'd hope any issues are dealt with.
Nvidia's developments with CUDA have been very good and show a lot of actual real-world use, unlike AMD's GPGPU developments so far. OpenCL will probably change this but that's how it is now.
One day, the aim will be that there is a single processor using fast internal Ram. This locks Nvidia out entirely unless Intel partners up with them and that looks unlikely. For AMD/ATI it doesn't matter as they do their own CPU solution anyway.
If it comes down to that level, I'd rather have an AMD/ATI solution Than an Intel one with Intel graphics.
It will certainly be an interesting turn of events if lots of companies start turning their back on NVidia.
=
Great post
AS I understand it nvidia and intel both have interlocked agreements with each other.
MAD mutual assured destruction.
So neither one will go all the way against the other.
And after the AMD SCANDAL WITH INTEL,everyone will play fair.
Comments
Makes you think, though the 210m isn't a motherboard, it does succeed the 9400m card, so maybe in the near future the MBPs will have a 210m motherboard and a 230/40 discrete card to replace the current setup.
The 250m gts doesn't show OpenCL compatibility on the Nvidia site, so it seems illogical. And an increase in power consumption seems silly since it would reduce the battery life claims Apple makes a big deal out of right now.
I was looking through the newly announced 200m series cards and only the 210, 230, and 240 support OpenCL and the 230/40 both have the same power consumption as the 9600gt they succeed.
Makes you think, though the 210m isn't a motherboard, it does succeed the 9400m card, so maybe in the near future the MBPs will have a 210m motherboard and a 230/40 discrete card to replace the current setup.
The 250m gts doesn't show OpenCL compatibility on the Nvidia site, so it seems illogical. And an increase in power consumption seems silly since it would reduce the battery life claims Apple makes a big deal out of right now.
Well, i am certain all G200 and G92 based Nvidia Product support OpenCL. So may be the page isn't updated yet.
The Reason why GTS 250M got my attention was its Performance / Watt. Which nearly double the currently Best Mobile Offering on the Market.
The GT240M will give 50% !!!! more performance then the current 9600M GT used in MBP at the same power usage!!!!
And GTS 250M will give MORE THEN 150% INCREASE in PEROMANCE of current 9600M GT. Using only 20% more power.
I cant wait till Apple get these update. It is a Huge Leap comparer to current offering.
The good news is that for us MBP users, the difference will be like night and day.
The problem is that these hot new GPUs aren't compatible with Arrandale, with its integrated GPU and memory controller. Using them would require Apple go to a 3-chip solution from the current 2-chip one. Right now the 9400m has an integrated memory controller, so Apple only needs 2 chips (CPU + GPU), but if Apple wants to work around the crummy performance of Intel's new integrated GPU (1/2 that of the 9400m) it needs an external controller in addition to the external GPU. The extra heat and cost will not be to Apple's liking. Intel deliberately designing their CPU to shut out the competition? Say it isn't so!
240 and 250 will have exactly the same type of interconnection as current 9600M GT (it uses PCI-Express x16 bus and dedicated GDDR3 memory onboard).
And only difference between current and future generation will be in Nothbridge that doesn't have IMC and IGP.
Apple now uses 3 chip solution: CPU + NB (which has 9400M GT and other stuff integrated) + 9600M GT
240 and 250 will have exactly the same type of interconnection as current 9600M GT (it uses PCI-Express x16 bus and dedicated GDDR3 memory onboard).
And only difference between current and future generation will be in Nothbridge that doesn't have IMC and IGP.
Apple now uses 3 chip solution: CPU + NB (which has 9400M GT and other stuff integrated) + 9600M GT
Since i just bought the 15 "3.02GHz 500g7200rpm 3 chip
design .you guys suck .
so can i switch OUT the shitty 9600 for the 240 ?
And are you saying all 3 chips work together on the MBP ?
240 and 250 will have exactly the same type of interconnection as current 9600M GT (it uses PCI-Express x16 bus and dedicated GDDR3 memory onboard).
And only difference between current and future generation will be in Nothbridge that doesn't have IMC and IGP.
Apple now uses 3 chip solution: CPU + NB (which has 9400M GT and other stuff integrated) + 9600M GT
Apple's low end laptops ( and the Mini) don't have a 9600m. Remember that Arrandale doesn't even have a Northbridge, it's all integrated on die. Since Larrabee is 45nm while Arrandale will be 32nm, my bet is the IGP resides in the same package as the CPU, but is not on die. Do you think Apple will willingly suffer 1/2 the graphics performance from Larrabee (in comparison with 9400m)? Maybe, but that will be a tough sell for its customers. That means an external IGP, but Intel claims Nvidia doesn't have the required license to integrate with the on-die IMC. While that remains in litigation, the only way Apple will be able to maintain graphics performance for products that don't have their own video memory is to use an external GPU along with an external memory controller. That's likely to be costly in heat, motherboard space, or more likely both.
Apple's low end laptops ( and the Mini) don't have a 9600m. Remember that Arrandale doesn't even have a Northbridge, it's all integrated on die. Since Larrabee is 45nm while Arrandale will be 32nm, my bet is the IGP resides in the same package as the CPU, but is not on die. Do you think Apple will willingly suffer 1/2 the graphics performance from Larrabee (in comparison with 9400m)? Maybe, but that will be a tough sell for its customers. That means an external IGP, but Intel claims Nvidia doesn't have the required license to integrate with the on-die IMC. While that remains in litigation, the only way Apple will be able to maintain graphics performance for products that don't have their own video memory is to use an external GPU along with an external memory controller. That's likely to be costly in heat, motherboard space, or more likely both.
Apple will likely use discrete G210M to complement Arrandale's IGP in low-end segment as well.
There is a rumor that Nvidia is out. http://www.semiaccurate.com
I can't imagine Apple reverting back to the poor performance of Intels chips so what is ip? AMD I don't think so. An internally built GPU, possibly.
Apple seriously needs to worry about Intels behaviour with respect to GPU producing third parties as it is not in their best interest. The risk is high that they would get locked into poor performing and buggy GPUs and be extremely limited in their system design abilities. I just see going all Intel as a serious long term mistake.
In anyevent has Nvidia been all that much of a reliability issue after this issue was identified. I'm just not hearing about excessive numbers of Macs coming back with GPU failures anymore. Considering the solid performance increases coming in the same power profiles what else is there for Apple to complain about?
As to internally produce GPUs I can see an arguement for this. Especially in the case of a GPU that is also an optimized OpenCL engine. The bad part is that it would be proprietary to Apple. The good part is that it's structure could be tightly tailored for Apples needs.
I guess time will tell here but I'd be very surprised to see Apple step away from Nvidia. The relationship seems to be firing on all cylinders right now. If it has gone bad it must be over something that is not public at the moment.
Dave
I guess time will tell here but I'd be very surprised to see Apple step away from Nvidia. The relationship seems to be firing on all cylinders right now. If it has gone bad it must be over something that is not public at the moment.
My bet is it's something very public: Nvidia's litigation against Intel. If Apple wants to get its new integrated-GPU products out on time and still use Intel's CPUs they must drop Nvidia. Litigation is so uncertain that Apple can't afford to place bets on who will win or when it will end, nor can they afford Intel refusing to ship them product, or put them on the back of the line instead of the front, because Apple continues to do business with Nvidia.
Not selling Nehalem licenses makes me assume that squeezing out third-party vendors in the integrated-GPU space was the entire point of the Arrandale design.
My bet is it's something very public: Nvidia's litigation against Intel. If Apple wants to get its new integrated-GPU products out on time and still use Intel's CPUs they must drop Nvidia. Litigation is so uncertain that Apple can't afford to place bets on who will win or when it will end, nor can they afford Intel refusing to ship them product, or put them on the back of the line instead of the front, because Apple continues to do business with Nvidia.
Not selling Nehalem licenses makes me assume that squeezing out third-party vendors in the integrated-GPU space was the entire point of the Arrandale design.
Are you sure that Arrandale will provide RAM interface for integrated videocard?
I'm pretty sure that using discrete GPU with dedicated memory is only possible solution regardless of chipset.
And I really hope Apple to use Intel chipset instead of nVidia's one combined with 210M for 13" (or AMD equivalent) and 240/250M for 15/17 (or AMD equivalent).
You'll get best of all worlds - minimum power consumption of Intel's iGP, maximum stability with Intel's chipset and still high 3D performance with powerful discrete GPU.
There is a rumor that Nvidia is out.
Seems like that guy has an issue with Nvidia himself and wants to see AMD/ATI do better. Apparently he posted something about Dell and Nvidia too, which hasn't come to anything.
The best way to get traffic to your site is to post a plausible rumor like that with no evidence that it took place and no way to prove it didn't but one that has important consequences.
The problem with the defective GPUs is quite far reaching for Apple as it could affect every single Macbook Pro shipped in the last 2-3 years but as mentioned above, there doesn't seem to be a huge return rate noted anywhere.
It's going to happen one day that Apple will drop external GPUs altogether, possibly when they hit 22nm Ivy Bridge in 2011. If Intel's chips weren't so terrible right now, I wouldn't mind. But they are, so for the time being, I'd hope any issues are dealt with.
Nvidia's developments with CUDA have been very good and show a lot of actual real-world use, unlike AMD's GPGPU developments so far. OpenCL will probably change this but that's how it is now.
One day, the aim will be that there is a single processor using fast internal Ram. This locks Nvidia out entirely unless Intel partners up with them and that looks unlikely. For AMD/ATI it doesn't matter as they do their own CPU solution anyway.
If it comes down to that level, I'd rather have an AMD/ATI solution than an Intel one with Intel graphics.
It will certainly be an interesting turn of events if lots of companies start turning their back on NVidia.
My bet is it's something very public: Nvidia's litigation against Intel. If Apple wants to get its new integrated-GPU products out on time and still use Intel's CPUs they must drop Nvidia. Litigation is so uncertain that Apple can't afford to place bets on who will win or when it will end, nor can they afford Intel refusing to ship them product, or put them on the back of the line instead of the front, because Apple continues to do business with Nvidia.
This really makes sense to me, however I still can't see Apple going back to Intel hardware for GPU's no matter how slightly improved they would be. It would be a very hard sell unless those iGPUs are a lot better than what many of us expect. That would mean good 2 & 3D support and functional video acceleration hardware. Or an architecture that can accelerate video without the need for special hardware. Intel just doesn't inspire me here.
Not selling Nehalem licenses makes me assume that squeezing out third-party vendors in the integrated-GPU space was the entire point of the Arrandale design.
This I'm very concerned about. It is something that needs proactive regulatory action if you ask me. If nothing else it will have a dramatically negative impact on innovation and advancement of the art.
Dave
I still can't see Apple going back to Intel hardware for GPU's no matter how slightly improved they would be.
Improved? Larrabee is supposedly 1/2 the performance of the 9400M. The 200 series just puts them even further behind.
Seems like that guy has an issue with Nvidia himself and wants to see AMD/ATI do better. Apparently he posted something about Dell and Nvidia too, which hasn't come to anything.
The best way to get traffic to your site is to post a plausible rumor like that with no evidence that it took place and no way to prove it didn't but one that has important consequences.
The problem with the defective GPUs is quite far reaching for Apple as it could affect every single Macbook Pro shipped in the last 2-3 years but as mentioned above, there doesn't seem to be a huge return rate noted anywhere.
It's going to happen one day that Apple will drop external GPUs altogether, possibly when they hit 22nm Ivy Bridge in 2011. If Intel's chips weren't so terrible right now, I wouldn't mind. But they are, so for the time being, I'd hope any issues are dealt with.
Nvidia's developments with CUDA have been very good and show a lot of actual real-world use, unlike AMD's GPGPU developments so far. OpenCL will probably change this but that's how it is now.
One day, the aim will be that there is a single processor using fast internal Ram. This locks Nvidia out entirely unless Intel partners up with them and that looks unlikely. For AMD/ATI it doesn't matter as they do their own CPU solution anyway.
If it comes down to that level, I'd rather have an AMD/ATI solution Than an Intel one with Intel graphics.
It will certainly be an interesting turn of events if lots of companies start turning their back on NVidia.
=
Great post
AS I understand it nvidia and intel both have interlocked agreements with each other.
MAD mutual assured destruction.
So neither one will go all the way against the other.
And after the AMD SCANDAL WITH INTEL,everyone will play fair.
I think.
just saying
9