Tag Archives: computing

WWII code-breaking computer goes head to head with modern pc

A legend reborn… or rebuilt at leastThe ten ‘Colussus’ code-breaking machines, thought to be the first modern digital computers, broke the code of many German communications through the second world war. To commemorate the work done by the codebreakers at Bletchley Park, Tony Sale has rebuilt one of the machines, which were broken down after the war for security reasons.

Now, with the rebuilt machine functioning, a competition is being run to decipher a German transmission similar to those seen during the war. The rebuilt Colussus will compete with a virtual version of the codebreaking program on a modern computer. Due to the single-minded nature of the Colussus, it’s closer than you may think. Sometimes a multi-purpose personal computer isn’t as good as a van sized monstrosity made up of 2000 valves.

[via BBC, image via picotech]

Green data centre uses 93% less energy than a normal collection of servers

Compellent combines efficiency with low energy use

Wow. Colour me impressed. Compellent Technologies produce network servers and data storage hubs. With the rise of the internet, the amount of power generated by servers and data has grown exponentially, with most internet servers needing vast quantities of power and cooling of the heat produced. Compellent have made a product that only uses power when the data is being accessed, using a load of technologies like Automated Tiered Storage, Thin Provisioning and Advanced Virtualization. This can cut the power usage of the company buying the data centres by up to a massive 93%. If every server used technology like this, a large chunk of every developed country’s electricity usage would disappear.

[via Treehugger, image via Compellent’s website]

Making the internet more like E-Coli

Does the net work in a similar way to the bacteria that makes us ill?There’s fascinating article on Discover today about a control theorist called John Doyle working on ways to improve internet speeds. He compares the structure of the internet to the E-Coli bacteria – both structures resemble a bow tie in the way they homogenize information or DNA into a small knot in the centre then spread them out to their respective destinations. With the oncoming prospect of RFID on most products and wireless nodes popping up all over the place, having an internet structure that doesn’t collapse under the weight of all the signals broadcast across it is essential. Doyle thinks that by letting computers use more information about internet traffic flow and speed, they can use the quickest route more easily, speeding up transmission of data by a huge amount.

[story via Discover magazine, image by sdbrown]

supercomputers that fit in your hand

Harvard image of tiny nanowireResearchers at a university in Scotland believe that thanks to ever-expanding research into ultra thin nanowires, supercomputers the size of matchboxes might not be more than a decade away. Nanowires, some 1000 times smaller than a human hair, have exhibited strange behaviour due to their small size but the scientists at the University of Edinburgh think they have worked out how to minimise it, leading to their bold prediction.

The department of Physics where I study in Bristol has a massive new nanoscience building nearing completion. The field is full of promising breakthroughs for micro-sized (and so less energy intensive) devices, especially in computing. Anyone hoping to build a palm held supercomputer may well use devices like the holographic nanoassembler coupled with high speed atomic force microscopy to put together such tiny machines. The holographic nanoassembler is especially fascinating research as it never touches the tiny particles, using lasers to manipulate the smallest of objects. Nanowires are also incredibly useful for solar panels, where current efficiency is limited by the large metal substrates that carry electric charge, which obscure some of the sun-collecting surface.

[Photo by Harvard University via IEEE Spectrum]

Moore’s Law to end in fifteen years?

microchip Gordon Moore has predicted the expiry of the "Law" that bears his name to occur within the next ten to fifteen years. Moore’s Law is a rule of thumb that states that the number of transistors that can be inexpensively placed on an integrated circuit doubles every two years (or thereabouts), and it has held up remarkably well since Moore coined it in the mid-sixties.

Indeed, this isn’t the first time Moore has sounded a death-knell for the Law, but as conventional electronics is inherently limited by the laws of physics, it’s plausible that it has to stop at some point. So what does this mean for the exponential theories of Singularitarians like Ray Kurzweil? Or will technologies like quantum computing pick up the ball before semiconductors drop it? [Via SlashDot][Image by oskay]

[tags]Moore’s Law, computing, electronics, futurism[/tags]