Intel has launched a new central processing unit on a postage stamp-sized silicone chip containing 48 processing cores and 1.3 billion transistors, known as the Single-Chip Cloud Computer.

This means the firm has created the equivalent of 12 quad-core processors in a single chip. Theoretically, each processing core on the IA32 processor could run a separate operating system. And this device contains all the resources usually needing several racks in a data centre within a single chip.

There are 24 ’tiles’, each one a dual-core processor with its own router, and together they form a mesh network able to communicate at 256 Gb/s. Each tile can also independently vary the frequency at which the cores operate. The tiles are grouped in fours, with each group able to run at its own voltage.

To reduce power consumption, the tiles can be split into differently sized islands running at separate voltages. The tiles in the mesh network can be varied to function from 25W to a 125W maximum.

The focus of modern competition in CPU production has shifted from processor speed to the number of processors which can be crammed into a single chip. American company Tilera and graphics chip maker Nvidia have already unveiled 100 core and 512-core processors.

Intel’s own 80-core processor from 2007 was less about practical application, and more about proving that it could be done. And the company has hinted that it could eventually fit 100 cores onto one chip.

What sets the SCC apart is the fact that it is based on Intel’s X86 architecture, making it possible for it to function in regular desktop computer systems like Linux and Windows.

Microsoft has already said it would exploit the capability of the SCC in the future, and Windows 7 can manage up to 256-core processors. However, the SCC announcement has sparked fears that Microsoft and other software sellers could become involved in complex per-core licensing, which could make life complicated, not to mention costly.

The microprocessor was produced thanks to research co-led by Intel Labs researchers in India, Germany and the US.

While Intel has not said when the SCC would be made commercially available, researchers and other interested parties will have access to it during the first half of next year, and more details will be announced at San Francisco’s International Solid-State Circuits Conference next February.

Chief Technology Officer at Intel Justin Rattner said: “Over time, I expect these advanced concepts to find their way into mainstream devices”. Can’t wait!

The new remote from URC is the first to feature haptic technology. This is where the touchscreen buttons emulate the feel of a real button to the touch, and make the clicking sound we all know and love.

“It [haptics] gives you an immediate, very satisfying sensation,” said Eric Johnson, VP of URC. “When you look at the rest of the remote too, boy it feels so good and looks so good.”  This breathtaking compliment, that is normally only reserved for a woman, was followed by a profound statement.  “[Noises for button presses] can grow tiresome”.  Now there speaks a wise man.

The technology took three years to develop, and the buttons have been designed to stand over a million clicks. The MX-5000 (left) is styled after the MX-6000(right).  It is believed that the MX-5000 will be easier to use because it has haptic technology. 

The MX-50000 also boasts:

• 2.7-inch LCD color touch screen
• WiFi (B and G) for metadata, control, integration
• Narrowband RF and IR for control
• Via URC base station, integrates with IR and RS-232 components
• Includes rechargeable battery and charging base

According to Mr Johnson, the MX-5000 is the “epitome of engineering”.  Though this maybe an accurate assessment of the design quality of the remote, at a cost of $1500, which is around £900, I feel most people will be sticking to the remotes that come with their DVD player.  Most of which have their own buttons, that are real.

A release date for URC’s MX-5000 remote, is yet to be decided upon.

Source:ce-pro

You may love the ultra-realistic graphics, and the sound quality may be unbelievable, but, essentially, the way you ‘talk’ during computer games remains in a 1990s time warp. Could all this be about to change? Some experts reckon so.

Kevin Warwick, Professor of Cybernetics at Reading University, points out that computer game characters remain pretty two-dimensional. Pre-scripted conversations destroy even the slightest illusion of talking to a ‘real’ person.

He adds that, although the technology is there to improve conversation, developers have concentrated their efforts on the visual side of games.

But now chatbots could make things very different. A general purpose chatbot could be able to talk about any subject the user can think of, with virtually no limits. This is a clear departure from the current situation, where there is a finite range of conversation topics, for example relating only to the game being played.

This is certainly a challenge for chatbots like Ultra Hal, which won creator Robert Medeksza the Loebner prize – for fooling interrogators into thinking they are conversing with a human

He admits the extent of the challenge of creating realistic Artificial Intelligence characters within an infinite range of potential conversation topics, but intends to use Ultra Hal within a computer game setting.

“We will demonstrate the technology in a simple game, where chatting with AI characters will be important,” he explains.

Essentially, Ultra Hal learns from previous conversations, and its AI-driven personalities can assist players on their quest. There is a massive database of words.

“But, if a players chooses to speak off-topic, the character could talk about anything” Medeksza adds.

To work within a game’s story, the chatbot needs to be fully integrated with the engine, and the game programmed to communicate different stages to the AI component.

Other designers are having the same thoughts as Medeksza. Developer Rollo Carpenter’s Jabberwacky has been learning patterns of conversation from online users for a decade.

And at this year’s Electronic Entertainment Expo, the trade event which showcases computer games, Microsoft demonstrated Milo, a virtual boy apparently able to have a natural conversation.

With games characters able to talk, learn and develop distinct personalities, and headsets increasingly commonplace for console users, the conditions could not be better for games with the latest chat technology.

Who knows, we may even come to love our games characters more than the real people in our lives.

Leading on from our article about paperless offices becoming reality, ASUS unveiled some remarkable technology at Computex 2009.

The idea, is that to transfer and work on files, rather than plugging in memory sticks and messing around with other portable storage, the work surface, or table, simply reads devices placed on it. These are then displayed, and using touch-screen technology, manipulated.

Is this ‘Seamless Experience’ the future? This video shows the idea in action:

Other than being a clever bit of fun at an exhibition, it does raise questions about green technology, and about how we will work in the future. How, and if the technology is fully utilised would depend on how well we could manipulate the data stored on the devices.

Could we print it out, add information to it, or delete information from it. Could we copy it, paste it. Change the font? If it was mathematical data, could we add something to an equation, or change a formula.

The more you think about what it would need to have, the more questions are asked.

New technologies emerge all the time, and many are consigned to exhibitions for their moments of glory. The feeling about this one is that it is little more than a clever bit of fun.

However with green issues being discussed at all levels, could there be enough push to get this one off the ground? I have to say though, if the information was displayed in a desk, what were the monitors all over the table for?

Gallery image: ndevil/flickr