TSMC 28nm
Hi guys,
with all the rumours present that Apple would be switching it's processors from Samsung to TSMC for the Apple A6, do you think Apple will use TSMC's 28nm plant? And thus, use TSMC's 300mm wafers for the A6? Will they use ARM-Cortex A9 or A15?
I hope they use both the A15 and TSMC's 28nm... But I don't know if ARM would have released the A15 in time for the iPad 3.
with all the rumours present that Apple would be switching it's processors from Samsung to TSMC for the Apple A6, do you think Apple will use TSMC's 28nm plant? And thus, use TSMC's 300mm wafers for the A6? Will they use ARM-Cortex A9 or A15?
I hope they use both the A15 and TSMC's 28nm... But I don't know if ARM would have released the A15 in time for the iPad 3.
Comments
Hi guys,
with all the rumours present that Apple would be switching it's processors from Samsung to TSMC for the Apple A6, do you think Apple will use TSMC's 28nm plant?
I would hope so, the current A5 would be crap in an iPhone as it draws to much power. This would be surprisingly bleeding edge on Apples part though. So in a way what I'm saying is that the next "A" processor might be specifically for the iPhone, call it an A5 Mk2 if you will. For the iPad the next processor needs to be significantly faster so A6 will be 28nm or an equivalent low power solution.
And thus, use TSMC's 300mm wafers for the A6? Will they use ARM-Cortex A9 or A15?
A15 is of limited use, long term Apple needs to go 64 bit. A15 is a kludge for servers as such I believe it would be a waste for Apple to pursue it. A9 will keep them for a long while, especially if you consider that they might hit 2GHz in a SoC drawing less than 2 watts flat out.
I hope they use both the A15 and TSMC's 28nm...
A big mistake in my mind. The only way I could see A15 fitting into Apples future is if they used it as a interim solution until 64 bit hardware can be had. To go A15 is to be stuck in 32 bit land. Now many will say that is strange who needs 64 bit addressing on a tablet, to that I would answer we are more than a quarter of the way to using up the 32 bit address space. Combine that with the extreme memory requirements of a Retina display for the iPad and the desire for 64 bit addressing becomes clear.
But I don't know if ARM would have released the A15 in time for the iPad 3.
I'm fairly sure it won't be released in time. No big deal, remember A9 can run faster and with more cores so we will still see good progress in CPU speed. However this isn't as important as adding in a GPU that can drive a Retina display with increased performance. Apple will have to focus real hard on other aspects of the A6 processor to get it to work well with the new displays and other goodies. We might even see integration of other hardware on the SoC in an effort to limit power disapation.
A15 is of limited use, long term Apple needs to go 64 bit. A15 is a kludge for servers as such I believe it would be a waste for Apple to pursue it. A9 will keep them for a long while, especially if you consider that they might hit 2GHz in a SoC drawing less than 2 watts flat out.
A big mistake in my mind. The only way I could see A15 fitting into Apples future is if they used it as a interim solution until 64 bit hardware can be had. To go A15 is to be stuck in 32 bit land. Now many will say that is strange who needs 64 bit addressing on a tablet, to that I would answer we are more than a quarter of the way to using up the 32 bit address space. Combine that with the extreme memory requirements of a Retina display for the iPad and the desire for 64 bit addressing becomes clear.
I'm fairly sure it won't be released in time. No big deal, remember A9 can run faster and with more cores so we will still see good progress in CPU speed. However this isn't as important as adding in a GPU that can drive a Retina display with increased performance. Apple will have to focus real hard on other aspects of the A6 processor to get it to work well with the new displays and other goodies. We might even see integration of other hardware on the SoC in an effort to limit power disapation.
Oh, my mistake. From what I've read, the A15 is the successor to the A9. My bad.
Oh, my mistake. From what I've read, the A15 is the successor to the A9. My bad.
A15 follows on to the A9 in a sense, however it is designed as a processor for servers. Info isn't really all that complete on the A15 but it seems like they are shooting for a power profile that is higher than a device like an iPad could use. 28nm might deal with that but it would be guessing at this point to say A15 would be viable in a tablet.
In any event the way I look at the situation is this, the next processor Apple needs to build will need to focus on features other than the CPU. Apple will probably go the AMD route by dedicating more die space to the GPU than to the CPU in A6.