Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
There NEEDS to be a reason to use this amazing stuff more than we have now

How about this: Processor uses less power because it takes way less time to calculate "email and flappyturd" as you put it? Or apps like ProCreate, Photoshop, iMovie and so on. It makes a difference for battery life if a processor uses 3 watts for a few seconds or only 0.8 sec - same thing for LTE in comparison to 3G (or less). So we do see the gains somewhere, don't worry about it too much :)
 
No roadmaps go beyond 3nm. The trouble, aside from increasing manufacturing challenges, is that quantum effects start to dominate and the transistors don’t behave like they used to at larger geometries. Modern transistor modes have hundreds of device parameters that attempt to track all the relevant physical parameters that affect their performance on modern nodes.

The width of a silicon atom is 0.2 nanometers, so we are talking transistors with features in the tens of atoms already. There is a real physical limit, even if the quantum effects weren’t in play. This is why there are efforts to find a replacement for silicon that allows circuits to switch faster. There are already materials out there, but they cannot be manufactured on the same scale and density as the current CMOS processes.

3DIC techniques tackle it from the energy per bit angle by making the interconnects closer together, making them easier to drive between interfaces, and thus, able to drive faster. Thermal management techniques will also help because heat has a negative influence on transistor performance, and thus, speed.

How about graphene? :)
 
Not to mention: this 7nm process will likely also be used in the x86 industry for AMD Ryzen CPUs which could easily outshine Intel’s offerings - Intel finally would be forced to innovate again with better technology coming from AMD and ARM.
Yeah or apple would buy amd and make theyr own processors again...
 
Yeah or apple would buy amd and make theyr own processors again...

Risky, but an inspiring line of thought. My only concern would be that AMD relies heavily on Global Foundries, and that could be problematic. GloFo meets AMD's current QC, but not if Apple were the customer. They have failed 2X to receive orders to fab Apple Silicon. :apple:
 
Clearly these TSMC dudes spend more effort on the engineering part than on slides and logo's, I like that :D

It would look pretty much the same if you went to our developers and asked them to make a slide for next boardroom meeting :)
 
  • Like
Reactions: sos47 and ravenstar
TSMC are seriously catching up with intel in terms of processing performance. Makes Intel look like they're sleeping in terms of the progress they are making. Not surprised Apple is rumoured to be switching macs to its own A-variant chips in the coming years (this would be a really bad idea to do now.. plus I like booting into Windows to play games... good luck doing that with TSMC's chips).
 
How about this: Processor uses less power because it takes way less time to calculate "email and flappyturd" as you put it? Or apps like ProCreate, Photoshop, iMovie and so on. It makes a difference for battery life if a processor uses 3 watts for a few seconds or only 0.8 sec - same thing for LTE in comparison to 3G (or less). So we do see the gains somewhere, don't worry about it too much :)

And then Apple can fit a smaller battery as the current 1 day is regarded as the good enough standard.

Trust me, if Apple thought, YAYYYY, now it's all using so little power, we can fit the batteries to give them 3 days (on average) of power, then I'd love it.

Sadly it seems that's not how their minds work. :(
 
The width of a silicon atom is 0.2 nanometers, so we are talking transistors with features in the tens of atoms already. There is a real physical limit, even if the quantum effects weren’t in play.

It's worse than that. There is space between the atoms in crystal, and the lattice constant for silicon is quite large at 0.5 nanometers.
 
It's worse than that. There is space between the atoms in crystal, and the lattice constant for silicon is quite large at 0.5 nanometers.
Yes, this is a good point. It will be even worse yet if you’re utilizing any straining techniques as well.
 
  • Like
Reactions: sos47
This is all starting to remind me of when apple switched from the ppc processor and switched to intel. This will likely result in apple dumping OS X and moving to an OS that can only run on these chips. Which, of course, lead to everyone having to purchase new versions of the software that they use regularly. It was a cluster!@#$ then and it will be again. Just my honest opinion.

Article Link: TSMC Details Technology Roadmap With Multiple Offerings to Benefit Future Apple Devices[/QUOTE]
 
7nm... 5nm... what happens when they run out of nanometers? :eek:

Remember grade 2 math? 0.9 nm, 0.8nm, 0.001nm?

Besides that, 1 Angstrom = 0.1 nm. 1 picometer = 0.001 nm. 1 femtometer = 0.000001 n. 1 attometer = 0.000000001 nm. And there's more prefixes where those came from.

I also remember when memory chips were sold with die sizes in the 160 micron range; the move below 100 microns (100,000 nm) was a major psychological milestone. So switching to a new prefix isn't even that big a deal.
 
  • Like
Reactions: sos47
the artwork probably is from 1994, TSMC has been around a long time and they just use the same decks... this presentation is for engineers after all. it's not general marketing.

is "dice" really the correct plural form of "die" in this context?

yes it is, that's the field-specific jargon. the die comes from a wafer that has been diced.

thanks for this post. i used to be an asic designer and miss seeing this stuff. last design i did before i pulled the handle was TSMC40LP. i remember in the original days of 40nm that leakage current was a serious problem, but they managed to figure it out. it boggles my mind that we're down at 7nm and yet leakage current is still a thing of the distant past. first design i ever worked on was a 130um gate array!
 
  • Like
Reactions: thisisnotmyname
Going with TSMC now also means that Apple has to stay with their INFO packaging process. Its a kind of lock-in which prevents Apple from going to outside OSAT's to get better deals. I guess this works fine while TSMC has the lead to be locked in, but can be quite a trap down the road.
 
this is future. i really want a desktop/laptop replacement. so, apple please enhance your apps for ios/ipads
 
No roadmaps go beyond 3nm. The trouble, aside from increasing manufacturing challenges, is that quantum effects start to dominate and the transistors don’t behave like they used to at larger geometries. Modern transistor modes have hundreds of device parameters that attempt to track all the relevant physical parameters that affect their performance on modern nodes.

The width of a silicon atom is 0.2 nanometers, so we are talking transistors with features in the tens of atoms already. There is a real physical limit, even if the quantum effects weren’t in play. This is why there are efforts to find a replacement for silicon that allows circuits to switch faster. There are already materials out there, but they cannot be manufactured on the same scale and density as the current CMOS processes.

3DIC techniques tackle it from the energy per bit angle by making the interconnects closer together, making them easier to drive between interfaces, and thus, able to drive faster. Thermal management techniques will also help because heat has a negative influence on transistor performance, and thus, speed.

I was designing CPUs for AMD and others back when we were going from microns to nanometers, and back into the 1990s. since 1994 I’ve heard them say every year that we were at the end of the road.

They always make assumptions. 2D trench, bulk silicon, etc. For example, I bet you are assuming FETs. There’s a lot of room between 3nm and 0.2nm if you are talking about the base thickness of a HBT (where lateral dimensions don’t really matter). (And if you design using CML you don’t have to worry about not being able to turn off transistors).
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.