Feeds:
Posts
Comments

Posts Tagged ‘HP’

IEEE’s Spectrum magazine has an excellent article about memristors and their history. This is an excellent overview piece written for a wider audience.  It clearly explains how the memristor came to be, why it matters, and what exactly it is.

It’s worth noting that Memristors have relation to cellular automata and neural networks, at least in originators.  Leon Chua is one of the main researchers behind Cellular Neural Networks and predicted the existence of memristors.  Also of note is the typical path of MIT, U of Illinois Champaign and Berkley – shared by many others working in similar disciplines.

Methinks the relationship between memristors and other cellular automata like theoretical models is much deeper than just the research instituations.  Wolfram mentioned the possibility of a parallel computing architecture based on CAs, perhaps the memristors plays some role in all of this.

Anyhoo, read the article.  Memristors will be significant in computing design.

Read Full Post »

Typically I am put off by analogies but consider…

Scientific symmetry demands that there should be a fourth fundamental circuit element to compliment resistors, capacitors and inductors. The fourth circuit, which would make it possible to build more complex electronic devices was theorized to exist by Leon Chua, a professor at the University of California at Berkeley, and called a “memristor” (memory resistor), that registers [remembers?] how much current flowed through a device.

Stan Williams, who heads up the Information and Quantum Systems lab at HP Labs, thinks that they have found it and how it works. – Nature, April 30, 2008 –

Forget all the sophisticated electronics and chemistry which should launch a new programming branch in a move away from silicon. Instead, imagine, if you will, that this discovery is analogous to how memory works neurologically. Sounds like blasphemy buy consider it for a bit. Without staining the logic, consider a memristor-like functioning of sets of neuorchemical – electrochemical flow rather than memory being a place or network of neural areas.

A memristor stores information based on changers of the resistance when an electrical current is applied. Memristors can have a high level of resistance, interpreted as a “1” in data terms, and a low level interpreted as a “0.” Thus, information [data] is electronically recorded (– absorbed – present, whatever new diction describes it) by controlling current. The memristor acts as a variable resistor that, through changes in its own resistance, reflects its own history, according to Williams of HP.

In an analogous scheme being suggested here for organisms, memories occur due to the flow and patterns of resistance changes in the 100 billion neurons in the brain factored by other billions of interconnections of neuro-electrical impulses that exist throughout the brain and central nervous system.

The amount of energy required to push around a very small number of atomic vacancies in a crystalline material to get memristors to work parallels what would be required to move electrical impulses throughout the nervous system net that is hypothesized to be what memories are. Williams said. “We can switch it [current] very fast, faster than we can measure” which is just what is experienced when evoked potential recordings are captured in neural recordings which makes them almost useless considering the number and speed of the differing neurotransmitters.

Memristor chips will be designed function like flash memory and retain data even after a computer is turned off, just like happens to some neural areas during sleep, conscious attention and the flow of moving in an out of the “zone” at work, play or when focused on any value-laden activities in life.

Just a thought.


This was written after an article in CNet.com www.news.com/8301-10784_3-9932054-7.html?tag=nefd.top

Read Full Post »