Researchers build first working memcomputer prototype.
38 replies, posted
[QUOTE=Swebonny;48140493]And apparently it takes exponential time to read the answer.[/QUOTE]
so basically it's worthless?
[QUOTE=Nebukadnezzer;48145316]so basically it's worthless?[/QUOTE]
I wouldn't say it's worthless since the whole memcomputing thing seems to be quite interesting and new. But the claim that it can solve NP-hard problems in polynomial time is extremely big. It kinda be equivalent to claiming that a cure to all cancers have been found.
Things like public-key encryption and many other encryption methods would be useless with these computers around. Getting the best solution to a problem would take place much faster and so on.
[QUOTE=Zombie_2371;48138581]Can someone with an understanding/background in this sort of stuff please explain in laymams terms what this really is all about?[/QUOTE]
One way to think of it is this:
Deterministic machines can only do one thing at a time. Thus given a problem that has many possible branches it must explore each branch at time. If a branch is bad (not the solution) it has to go back to one of the other branches and start over.
Non-deterministic machines can explore each branch at the same time. If a branch is bad it doesn't matter since it is already exploring the correct branch at the same time. Another way to think of it is that for every branch the machine encounters it ends up making a new deterministic machine just for that branch.
A classic example is the [URL="https://en.wikipedia.org/wiki/Travelling_salesman_problem"]traveling salesman problem[/URL]. Given a set of cities and the roads between them, find a route for a salesman to take such that 1) he visits every city exactly once and 2) he takes the shortest path. Hopefully you see how this problem creates those branches at each new city.
[editline]7th July 2015[/editline]
[QUOTE=Swebonny;48140493]And apparently it takes exponential time to read the answer.[/QUOTE]
Where's it say that? Edit: nevermind, googled for it.
[QUOTE=Amiga OS;48139777]Already a thing.
[img]http://www.jklaptopcare.com/images/lapImages/IBMT61.jpg[/img][/QUOTE]
Hey, those things are great. Running Debian on one and it can barely handle Source engine stuff.
Is that whole P=NP Millenium problem relevant to this?
-snip-
[QUOTE=Amiga OS;48139777]Already a thing.
-picture of a cheap Lenovo T61 laptop-[/QUOTE]I'm using one of those right now! I love it. It was cheap and it's managed to not explode violently, despite being surprisingly flexible.
[editline]8th July 2015[/editline]
I play Dwarf Fortress on it :]
[QUOTE=thrawn2787;48145719]One way to think of it is this:
Deterministic machines can only do one thing at a time. Thus given a problem that has many possible branches it must explore each branch at time. If a branch is bad (not the solution) it has to go back to one of the other branches and start over.
Non-deterministic machines can explore each branch at the same time. If a branch is bad it doesn't matter since it is already exploring the correct branch at the same time. Another way to think of it is that for every branch the machine encounters it ends up making a new deterministic machine just for that branch.
A classic example is the [URL="https://en.wikipedia.org/wiki/Travelling_salesman_problem"]traveling salesman problem[/URL]. Given a set of cities and the roads between them, find a route for a salesman to take such that 1) he visits every city exactly once and 2) he takes the shortest path. Hopefully you see how this problem creates those branches at each new city.
[/QUOTE]
Thanks! That helps explain it's functionality, but what are it's practical uses? How exactly can we see this applied in day to day life? What benefits will it net us?
Sorry, you need to Log In to post a reply to this thread.