Welcome to Laser Pointer Forums - discuss green laser pointers, blue laser pointers, and all types of lasers

LPF Donation via Stripe | LPF Donation - Other Methods

Links below open in new window

ArcticMyst Security by Avery

Memristor






Joined
Oct 14, 2008
Messages
233
Points
0
Interesting...It's almost like a vertically stacked, self-contained MOSFET.

This looks like it could eventually make its way into RAM/ROM storage, but I have so many questions about it regarding quantum effects, read/write margins, and more that would bore the socks off of most everyone here.
 
Joined
Sep 5, 2008
Messages
164
Points
0
Its advanced, i beleive he said a single memristor's resistance can represent more than an entire byte's worth of transistors, which makes sense if the resistance changes are large enough to have significant resolution. and something fast enough to decode that value into binary without creating a bottle neck. so was he saying it could work kind of like this?

1mohm = 00000001
and
130mohm = 01000010
and so on.

the proportional value of a single memristor is equal to the binary output of what would normally require multiple transistors?

I'm not saying it has to be that linear by any means, but it makes more sense as an example that way.


I wonder if they are volatile, or if they remain in the set state when power is lost? that would make for a new feature in ram! you could retain all information in ram even after a power cycles. maybe we'd see little ram partitions become a fad again.

Either way, SETs will put this kind of stuff to shame when they are perfected. but i think we'll see fusion power generation before that ever happens.  :(
 
Joined
Oct 14, 2008
Messages
233
Points
0
He said "replace ten transistors"..which means he's talking about replacing a single SRAM cell (one bit of memory) that uses between 6 and 10 transistors. After enough cycles, the sensing technology will perform destructive reads, meaning this is more suitable as a DRAM replacement than a SRAM replacement.

These memristors will be too small and will be driven too fast to perform discrete measurements of their value. You'll be able to register a 1 or a 0, and that's about it.

I hope they figure out how to minimize electron drift - a design like what they propose would definitely be volatile.
 

treb76

0
Joined
Oct 1, 2008
Messages
283
Points
0
They specifically say that it is non-volatile. If you remove the voltage from the memristor, it retains its resistance. These have the potential to lead to the instant on computer or no data loss when your laptop dies or a power outage hits your desktop. But I think they tell you all that. :p

There're also some videos on youtube from a symposium about the memristor. They give more information about the theory/math and applications for such a part.
 
Joined
Jul 27, 2007
Messages
3,642
Points
63
Definitely an interesting video, and yea from the video it would be non-volatile. Write speeds? bit resolution? That will be for future fabs and processing to figure out :p
 




Top