In about 10 years, Apple has multiplied by 19 the number of transistors in its mobile processors. It corresponds roughly to a steady rate of improvement of 34% per year on the number of transistors, or a doubling every 2.5 years. In real dollars, an iPhone has roughly a constant price: the price tag of a new iPhone increases every year, but it does so while tracking the inflation. Thus you are getting ever more transistors in your iPhone for the same price.
| processor | release year | transistors |
|---|---|---|
| Apple A7 | 2013 | 1 billion |
| Apple A8 | 2014 | 2 billions |
| Apple A9 | 2015 | 2 billions |
| Apple A10 | 2016 | 3.2 billions |
| Apple A11 | 2017 | 4.3 billions |
| Apple A12 | 2018 | 6.9 billions |
| Apple A13 | 2019 | 8.5 billions |
| Apple A14 | 2020 | 11.8 billions |
| Apple A15 | 2021 | 15 billions |
| Apple A16 | 2022 | 16 billions |
| Apple A17 | 2023 | 19 billions |
That’s an interesting observation. It would be a good addition to track process level change (nanometers) but I imagine it follows that well. Chip-size might be another nice one.
Plotting the data seems to suggest that transistor count increase is slowing down:
Note that the y-axis is logged to show multiplicative increases as a linear trend. I was going to say that 2018 was the end of exponential growth but looking at the gradient doesn’t show much happening then. Presume it’s being limited by the increasing difficultly of engineering these bigger devices.
there was an image in my comment but it seems to have been stripped out, see https://i.stack.imgur.com/lnQMw.png for the plot.
[feel free to delete this comment the original image get whitelisted)
I see three different trend lines. One to 2018, and then two more each with less extreme slopes consisting of two years each. Unknown why this is.
Moore’s law still at its works, although it may have paced down a bit.
I wonder if we can reach sub-nanometer resolution with optical technologies, as it is already deep in x-ray range. Is that finally the end of silicon-based computing as we know it, which has been announced for long but engineers kept teaching us otherwise so far?
Are there any serious contenders on the horizont to dethrown current serial computing technology? Maybe from nanotechs? Also it seems quantum computing will not hold up to the high hopes many marketeers have promised (thats at least what some physicists say), not to mention the algebraic issues around it…
billion. not billions.
I think it is “how many billions” and not “how many billion”:
https://ell.stackexchange.com/questions/136249/how-many-billion-or-how-many-billions
The English language is an idiosyncratic mess, but I’d interpret the (overly terse!) comment as referring to the “billions” appearing repeatedly in the table rather than the title.
My understanding is that if billion is used as a noun, then it takes plural form.
In this instance, I want to know how many billions there are.
Yup, think I agree on use of adjective vs noun, but reading your table pedantically I’d borrow the “transistors” from the heading and interpret, e.g., “8.5 billions” as being an abbreviation of “8.5 billion transistors”. Hence the use of “billions” in the rows incorrectly applies a plural to an adjective.
That said, I’d probably move the billions from the rows up into the heading as it’s repeated in every row, maybe as “Transistors (billions)”. The author of the latex package “booktabs” influenced my thinking about this a lot.
I didn’t notice this when I was reading the post when it was published, only when I got notifications about subsequent comments. What I did notice was the varying precision in the numeric values, e.g. the first few rows have one significant figure while later you have up to three significant figures. Presume this is just a limitation of the data you received and I could get more accuracy if I cared.
The source is wikipedia and I kept the accuracy that they provide. If you know of a more precise source, please share.
But with bloated software, so no gain for the user.