In which I post about a TL9 solar system

[Say, it isn't that bad!]

Originally Posted by maximara

I think that is more due to the limitation on what you can do with a CISC based chip. The iPhone uses a 7nm ARM chip (A12) and the A14 (a Bionic-based, 5 nm 12-core CPU) is supposed to show up in a Mac in 2021.

Though we are going to hit the quantum tunelling wall soon even with functional prototype 3 nm chips due out in 2021. I'm not sure if the gate-all-around FETs for 2nm chips is going to pan out and that may be the limit with RISC based instruction sets.

CISC or RISC doesn't change the fundamental size of transistors; although Arm being a (relatively) modern design with little baggage does help decrease the number of transistors needed.

Ultimately, though, Apple changing from x86 to Arm may be an indication that they are attempting to get around the soft wall of transistor size as much as practical.

[maximara]

Originally Posted by Say, it isn't that bad!

CISC or RISC doesn't change the fundamental size of transistors; although Arm being a (relatively) modern design with little baggage does help decrease the number of transistors needed..

I was thinking more along the number of transistors to do the some thing.

"The RISC ISA emphasizes software over hardware. The RISC instruction set requires one to write more efficient software (e.g., compilers or code) with fewer instructions. CISC ISAs use more transistors in the hardware to implement more instructions and more complex instructions as well." - RISC vs. CISC Architectures: Which one is better?

Though this is no few lunch: "The RISC architecture will need more working (RAM) memory than CISC to hold values as it loads each instruction, acts upon it, then loads the next one."

This explains why the A12z WWDC 2020 Special Event Keynote Mac had 16 GB of RAM compared to 8 GB one tends to se in the i5 and i7 low end Macs.

IIUC one of the issues Intel is having at going below 14 nm is heat and in space heat is a problem as the only option for cooling things off is radiation. The x86 architecture hasn't exactly been cool (I still remember the jokes about the Pentium space hear with the very first PC to use that chip)

Originally Posted by Say, it isn't that bad!

Ultimately, though, Apple changing from x86 to Arm may be an indication that they are attempting to get around the soft wall of transistor size as much as practical.

If that is the case then Apple planned for the really long game as the A4 (2014) used a 45 nm process while its x86 (Haswell-DT) used a much smaller 22 nm process. The fact that the WWDC mac handled x86 optimized code as well as it did shows how far software has come (the RISC PowerPC chip's emulation was annoyingly slow)

[Anaraxes]

Originally Posted by Say, it isn't that bad!

Ultimately, though, Apple changing from x86 to Arm may be an indication that they are attempting to get around the soft wall of transistor size as much as practical.

Two reasons, really:
(1) Making their own processor so they make more money on the hardware. ARM's long been in the business of providing processor cores to people doing their own ASICs. They don't sell a lot of standalone processors.
(2) Power efficiency. Intel's made a lot of improvements in that metric, but x86 has always been more about flagship processing power. ARM, on the other hand, has a customer base of people doing embedded devices, as well as mobile ones. To that market, power efficiency and low cost generally matters more than top-end performance. Apple laptops need to save power more than they need maximum MIPS, as do iPhones.

ARM doesn't make transistors any smaller than anyone else. In fact, transistor size comes from the chosen process; ARM's stuff is one level higher up.