For example i notice that the latest chips are 3 nm & below, & over the last decade i’ve seen it shrink little by little. What is it about this process that more money needs to be poured into each die shrink, & why couldn’t we just jump from 90nm to 3nm instead pf 65nm etc etc?
In: Engineering
Isn’t this the case with any technology ?
why did it take decades for cars to go over 400km/h ? why didn’t we get 1Gbps internet in 1996 ? why did we have to have full HD before 4K ??
Not a lot of explanations here that a 5 year old would understand, here’s mine.
Imagine trying to draw a very detailed house blueprint with a very fat marker – the lines would be too thick and would all blur together, and the blueprint would be unreadable. You need a pencil with a sharp point. A 90nm chip fab uses a “90 nm marker” to draw the parts of the chip, and if you tried to draw 3nm parts, the lines would all overlap and it wouldn’t work.
Making the “marker/pencil” sharper is technically very challenging – no matter how much you sharpen the pencil, it won’t make lines thinner than, say, a hair, so if you want to go smaller you need a whole different way of drawing lines in the first place. Each new generation of fabrication technology is like a new, sharper kind of pencil, and we’ve got that pencil so sharp that we’re hitting the limits of what’s physically possible.
I want you to build a doghouse. You have never done it before. I give you materials and tools and tell you to have at it.
Your first attempt takes a long time, looks like crap, will probably collapse on the dog, and wastes a lot of material.
I tell you to keep trying.
Your fifth attempt probably looks a lot better, is more structurally sound, and doesn’t waste too much material.
By the 10th attempt you’ve really got it down. You only use as much as you need, it looks good and is solidly built, and takes a reasonable amount of time. From this point out every one you build is identical to the one before and takes the same amount of time.
Now I tell you I want a two story house for my cat. You take everything you’ve learned and apply it, but its a slightly different process and you still make mistakes. It probably takes you 5 attempts this time to really get it figured out.
This is pretty much the process you go through when building anything. Its incremental progress. If you try to go directly from A -> E and skip all the in between steps, the chances of E failing, or taking a lot longer than you expected, or costing a lot more, are much higher. This also means that the only product you have to sell is A, until you are able to make E work. Meanwhile your competitor has introduced generation B and C and your sales on A are plummeting while you invest everything in making E work.
If instead you go A, B, C, D, E your chances of succeeding at each step are much higher, and you can do it faster and at a lower cost. And then you can put B on the market and sell it while you start to develop generation C,D,E.
I’m not familiar at all with the marketing stuff and only passingly familiar with the manufacturing part.
You can conceptualize the idea from a ton of different disciplines, though.
Consider someone wanted to make the world’s smallest car that performs identically or better than a normal car. Ignore all the nerd physics stuff like wind resistance, this is a 9th grade class where nothing matters unless I say so.
Starting with a normal car of today, you would start to come up with ideas about designs that are more compact.
It’s not long before you find the best layout and the car can’t get any smaller. You need that space for the engine. You put your best guys on it and they find a way to make an engine 5% smaller, so, you put in some effort and your car is now even smaller!
But, once people saw your new engine, they realized that the design could be generalized, so now you have a smaller battery, transmission, and someone came up with an even smaller engine. Put it all together and now your car is way smaller than it was before, but each part only made it a little smaller along the way.
Nothing technically would have stopped you from saying you wanted the smallest car and then doing it all at once, but it’s profitable to release things along the way and the market is REALLY good about pushing innovation. Maybe it would have taken you 10 years, but everybody together got it done in 3 because they were all competing with and learning from each other. This is one of the things that people who champion diversity like to point to: different perspectives can be HUGE for innovation.
Tldr: There was no point in time where people didn’t want any processors, so they are released continuously with small improvements rather than infrequently with larger improvements.
Latest Answers