Answer the question
In order to leave comments, you need to log in
Why not come up with a decimal computer?
I want to warn you right away not to write about zeros and ones, on / off. etc.
The question is rather that ten states could be stored in memory instead of two. Accordingly, memory access would be faster. Previously, when everything was in its infancy, it might have been difficult to implement technically. But now, with the current technical processes, is it really not possible to do this?
I'm not an engineer or an electronics engineer, so maybe I didn't quite correctly ask the question. But, nevertheless, the question arose.
Answer the question
In order to leave comments, you need to log in
It will worsen ... noise immunity or something. I'm not talking about the computers themselves, but about the methods of transmitting and storing information. Look, when transmitting data, there are now two states - 1 and 0. Let's assume that we transmit data by electricity (not optical fiber, but ordinary twisted pair). Roughly speaking, 0 is not the absence of voltage, it is a voltage from 0 to 1 volt (a number from the bulldozer). This covers measurement errors, static interference and other parasitic muck. Accordingly, from 1 to 5 volts is 1. Now, in the case of the decimal system, the scale narrows - 0 is from 0 to 0.5 volts, 1 is from 0.5 to 1, etc. Such a system will be less immune to interference and more demanding on the final signal meters.
Data transmission over the network in this case is just an example. This all applies to storage and transmission in any form.
The number system based on the natural logarithm - the number e, which is approximately equal to 2.71 - has the highest storage density.
This number is between two and three, but ternary logic is more expensive to implement than binary logic - if only because the transistor operating in linear mode generates much more heat than working key.
Your idea that supposedly memory access could be faster is at least not justified.
How do you physically imagine 10 states of one transistor, open, closed, ajar, and so on, or how do you imagine 10 states of disk magnetization, magnetized, slightly magnetized, even weaker magnetized ... there was once a ternary system, but they refused because haemorrhoids. In general, it is more difficult to implement at the physical level with the level of transistor logic that exists now, and why? A decade is only more convenient for a person because of the number of fingers, the decimal system has no other advantages.
The opportunity is there, but why? And it won't work any faster. because the signal has only two states - yes and no - 0/1.
Eniac replaced the decimal computer precisely because decimal is slower than binary. So binary is even more reliable, even on lamps. And information is able to store and process more. In the days of Eniac, this was not so noticeable, but now a decimal supercomputer will be inferior to an ordinary cluster server, and a desktop machine will catch up with it not in 10 years, but in 2-3.
It will be expensive and inefficient.
Ternary and analog computers didn't catch on.
https://ru.wikipedia.org/wiki/%D0%9A%D0%BE%D0%BC%D...
well, it's hard to imagine an element of the system, an element of the universe that holds ten states and can change them instantly. moreover, it is necessary that this element can be practically implemented at the technical stage of civilization - bicubic technologies, neurotechnologies, pn junctions, subatomic molecular particles, and so on .... IMHO, all this is not suitable. spin-electron - and that one can only have two directions, as well as a p-n-transition (n-p-transition), neurocomputers have the same biqubit structure - it seems there are three states .... I don’t remember exactly, but something something like. but for 10 or so, it’s complicated, and probably not why.
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question