Behind the scenes at Labs
Showing results for 
Search instead for 
Did you mean: 

For computer chip innovation, instant isn’t fast enough


1635 x 313 v2_abbreviated.jpg


In his latest lecture at HP, Professor Leon Chua noted that, as computers get faster, the time we spend waiting for computers to boot up becomes less important. Even if your computer shuts off suddenly, it takes only a few seconds to get back to what you were doing, and all your files are usually backed up automatically. This makes it seem like there isn’t much more to improve in current computer chip technology.


Here’s why. Most of modern computing relies on flash storage, a way of holding memory. But this method operates much too slowly to keep up with modern central processing units, so memory has to be stored and processed in different locations. This is why booting up a computer takes a few seconds instead of zero seconds.


L5_IMage.jpgWhile waiting for a computer to boot up is a minor inconvenience, it is still a symbol of how much farther there still is to go in the field of computer chip technology because of what it means about the fundamental architecture of computers.


Chips have been so packed with transistors that just adding more to increase processing speed is a strategy with a sell-by date.


Imagine instead there were a kind of computer chip that was dense enough to hold all of your information and fast enough to keep up with your processor. A lot more memory could be stored on one chip, and it wouldn’t need to be retrieved in order to complete an operation. This would have dramatic effects on the invisible processes behind all computer use. Issues like widespread energy conservation and security could be solved by radically rethinking how computer circuits should work.


This chip already exists. It’s called, Chua explained, a Memristor.


In his third of a 12-part lecture series at HP’s headquarters, Chua explained the symmetries and duals between all four types of fundamental circuit elements, and then how to recognize the characteristic behaviors of a Memristor as opposed to resistors, capacitors, and inductors. He also described how to classify different types of Memristors according to their behavior. 


Chua predicted the existence of the Memristor in 1971, and an HP Labs team, led by Senior Fellow Stan Williams, demonstrated a practical version of the device in 2008.


Memristive behavior can be found in many biological and physical systems, according to Chua. He demonstrated Memristor behavior in carbon-arc devices, devices first demonstrated by Sir Humphry Davy in 1802!  Much like the effects of Einstein’s theory of relativity, it took a new model to allow us to see what was always there.  


Professor Chua’s lecture is part of a 12-part lecture series taking place at HP’s headquarters at 3000 Hanover Street, Building 20 - Auditorium, Palo Alto, CA, 94304.


As a special addition to Professor Chua’s lecture, he is personally offered a prize of $1,000 to whomever can correctly answer this question, based on the topics he has discussed in his lecture series:


The Hamiltonian of a mechanical system (H) is the sum of the kinetic (T) and potential energy (V) and because of the conservation of energy, the sum total is conserved: H = T + V is a constant.  Think of the play between kinetic and potential energy in a pendulum, all potential energy at the top of the swing, all kinetic energy at the bottom of the swing, but the sum is constant.


As we examine the circuit of an inductor and a Memristor with the simplest relationship between flux and charge (φ = q + q³/3), we can derive an equation that looks a lot like the Hamiltonian:


H(q,i) = i + φ = i + (q + q³/3)  


When these systems are measured, H is being conserved in this circuit, but what is the physical interpretation of H?  What is being conserved?


At the beginning of his lecture, Professor Chua introduced us to the symmetries in the formulas and duals (resistance is the dual of conductance, capacitance is the dual of inductance). He ended it by closing the loop (pun intended) with this Noether’s first theorem: Every symmetry has a corresponding conservation law.  In the case of the pendulum, symmetry of time yields conservation of energy.


You can watch this lecture on HP on demand.





0 Kudos
About the Author