Wednesday, February 14, 2007

Start-up demos quantum computer

Michael Kanellos at CNet.com discusses D-Wave Systems plans to commericalized quantum computing in the months ahead.

******************

About a year from now, banks, pharmaceutical companies and other large institutions will be able to rent time on a computer that calculates by studying the behavior of a niobium atom, according to D-Wave Systems.

The Canadian company on Tuesday gave a public demonstration of Orion, its quantum computer, at the Computer History Museum in Mountain View, Calif. D-Wave said it is going to try to sell computing services to corporate customers in the first quarter of 2008.

Quantum computers, which researchers have experimented with for years but which haven't yet existed outside of the laboratory, are radically different than today's electronic computers. D-Wave's computer is based around a silicon chip that houses 16 "qubits," the equivalent of a storage bit in a conventional computer, connected to each other. Each qubit consists of dots of the element niobium surrounded by coils of wire.

When electrical current comes down the wire, magnetic fields are generated, which, in turn, causes the change in the state of the qubit. Because scientists understand how niobium will react to magnetic fields and calculate the exact pattern and timing of the magnetic fields created, the pattern of changes exhibited by the niobium can then be translated into an answer that humans can understand.
D-Wave

"The qubits behave according to a certain set of rules," said founder and Chief Technology Officer Geordie Rose, who likened quantum computing to trying to decipher the language of atoms. "Quantum computing is the translation of those laws into a format that we can take."

Ultimately, D-Wave's computer is an analog computer, according to Alexey Andreev, a venture capitalist at Harris & Harris and an investor in D-Wave. Answers to programs run on the computer come in the form of a physical simulation. Answers to problems in digital computers are essentially mathematical solutions.

Because of its inherent properties, D-Wave's computer is optimized for running complex and oftentimes consuming simulations--for example, what happens when different variables are changed in an ornate financial model, or how different proteins interact with various synthetic, simulated pharmaceuticals. The system also could be used for nonscientific research such as searching patent databases for matches and overlap of intellectual property.

"We view these machines as probability distribution generators," Rose said. "We want to build an actual physical embodiment of a hard math problem."

Right now, Orion is a "proof of concept," a demonstration of what the final product could look like. At the demonstration, Rose had the system come up with answers to Sudoku problems and, in another demo, seek out similar molecules to the active ingredient in the drug Prilosec in a chemical database. The computer found several molecules that shared similar structural elements with Prilosec, but the molecule that matched it closest was the active ingredient in another drug called Nexium. Plucking out Nexium demonstrated the system's accuracy, the company said. Nexium is actually a mirror image of the molecule in Prilosec that AstraZeneca invented to extend its patents.

In another example, he ran a seating chart program where each guest had particular seating requirements. (Cleopatra could not sit next to meat eaters. Genghis Khan eats meat, and so on.) The system came up with a seating plan with a minimum number of violations of protocol.

The computer itself--which is cooled down to 4 millikelvin (or nearly minus 273.15 degrees Celsius) with liquid helium--was actually in Canada. Attendees only saw the results on a screen. Still, it was the largest demonstration of a quantum computer ever, Rose said.

By the end of the year, however, D-Wave will have a 32-qubit system. It plans to begin to rent out time on its computers to corporate customers in the first quarter of next year, said CEO Herb Martin. Customer won't have to learn special programming techniques or other tricks to take advantage of the service; sending a problem to D-Wave will be similar to outsourcing it to any other company. Later, D-Wave may lease or sell computers, Martin added.

By the second quarter of 2008, the company plans to have a 512-qubit system, and a 1,024-qubit system is expected by the end of that year.

Quantum computers, Martin emphasized, will not displace digital computers. Instead, they will serve as co-processors for large problems.

But is there a market for renting computing cycles? Sun Microsystems a few years ago opened up a server farm for hire for chemical and pharmaceutical companies. It has found few takers.
D-Wave's appeal differs in that its computer will be able to solve much larger problems than companies are currently able to tackle, said Steve Jurvetson, a partner at Draper Fisher Jurvetson and an investor in D-Wave. Many medical outfits actually limit the scope of their research to fit the existing computational abilities. D-Wave has 100 patent applications filed, and 35 have been granted.

However, how larger systems will behave remains unknown, Rose said. D-Wave has engineered its chip so that the qubits are insulated from noise and other factors, and he has confidence that the number of qubits can be increased, "but we could be wrong," he said.

Another distinct advantage that the computer will have comes in energy consumption. Niobium is a superconductor and, thus, does not radiate heat. The chip itself requires only a few nanowatts.

The refrigeration unit consumes the most power at 20 kilowatts, which is still small compared with most server farms. Expanding the number of qubits on the chip will not require massive increases of refrigeration, Rose added.

Even with the explanations, quantum concepts can be a little tough to digest. Rose reminded the audience that humans consist of atoms that first appeared in a supernova billions of years ago. Trying to understand those atomic interactions that lead up to the present is at the heart of quantum computing.

"When I went to school, they didn't teach quantum mechanics," Martin said. "Newton was my boy."

Another Quantum Leap at IBM

Here's the lastest from the labs at IBM. Sounds like another quatnum leap to me...

**********************

IBM Pares Speed Gap In Memory Circuitry Design for Data Storage
May Be in Use Next Year;
Performance Could Double
By DON CLARK
February 14, 2007; Page B2

International Business Machines Corp. is claiming a breakthrough in developing circuitry to store data on future microprocessor chips.

The big computer maker said its approach -- based on exploiting the most widely used memory technology in a new way -- could triple the data stored on a typical microprocessor with a resulting doubling of computing performance.

"We think this is the next big thing in getting more system performance," said Lisa Su, IBM's vice president of semiconductor research and development.

Microprocessors, the calculating engines for computers, increasingly come with storage circuitry to minimize the delays associated with fetching data from external memory chips. This "cache memory" typically uses a kind of circuitry used on chips called SRAMs, or static random-access memories.

SRAMs are fast but require six transistors to store a single bit of data. The more widely used chip known as dynamic random-access memories, or DRAMs, only need one transistor and another component, a capacitor, to store a bit. But DRAMs, though they can store more data in a smaller space, have generally been considered too slow for cache memory.

IBM researchers are discussing their progress in closing the speed gap at a conference in San Francisco today.

Exploiting a manufacturing technology called silicon-on-insulator, the company has developed unusually fast DRAM circuitry for use as cache memory. Subramanian Iyer, a director of IBM's manufacturing-process development, estimates it takes 1.5 nanoseconds -- or billionths of a second -- to fetch data from its enhanced DRAM technology, compared with 10 to 12 nanoseconds for conventional DRAMs and 0.8 to 1 nanoseconds for SRAMs. Mr. Iyer said three times more data can be stored in the same amount of space by switching from SRAM to DRAM circuitry; he expects the technology to be incorporated on microprocessors that will be manufactured next year using a new production process.

Such benefits could help IBM's Power microprocessors in a performance race with chips from Intel Corp. and others. But Shekhar Borkar, the director of Intel's microprocessor technology lab, said extra manufacturing costs associated with using DRAM circuitry could outweigh the benefits.

IBM is a technology partner with Advanced Micro Devices Inc., an Intel rival that could benefit from the computer maker's memory research. Meanwhile, other alternatives to SRAM for cache memory are also being studied.

Innovative Silicon Inc., a start-up, has been promoting a technology it calls Z-RAM that stores data using a single transistor. Its licensees include AMD.

An AMD spokesman said it is "evaluating a number of new and emerging technologies" for cache memory.

Takin’ care of business, rather badly

Kevin Maney discusses the state of the online music business in his weekly USA Today column.

*******************

The online digital music business stinks.

ITunes, Rhapsody, Zune Store, Napster — you name it. They're all failures.

The hype has people believing otherwise. Bloggers, tech writers and your friends who know more about computers than you do shout that iTunes is the best thing to happen to music since the microphone. Or maybe psychedelic drugs.

But it's just not true. Nearly six years after the introduction of iTunes and the iPod, online music has failed to interest the vast majority of the world's music consumers. Which is no doubt why Steve Jobs recently called for an end to copy-protection software on digital songs. Something has to change, or iTunes and its ilk will never break into the mass market.

Jobs admitted that iTunes' penetration has been weak. In his discussed-to-death essay, "Thoughts On Music" — posted a couple of weeks ago on Apple's website — Jobs noted that only about 3% of songs on a typical iPod are bought on iTunes. The rest are either ripped from CDs and transferred into iPods, or illegitimately downloaded for free off file-sharing sites such as Kazaa or eDonkey.

The reality for iTunes might not even be that good.

In a report released in December, Forrester Research said it did a strenuous, independent analysis of iTunes purchases. It found that just 3.2% of all "online households" — homes that have computers and Internet connections, a subset of all homes — made an iTunes purchase over a one-year period.

About 10% of buyers purchased just one track during the entire year. About one-quarter of buyers spent $5 or less for the year. Most iTunes users, Forrester says, own fewer than two CDs' worth of iTunes music.

Really, it's as if tens of millions of people each had a big honkin' refrigerator, and put a quart of milk in it a few times a year.

Worse for Apple, Forrester found that the number of monthly transactions per iTunes household was declining in 2006. "It is too soon to tell if this decline was seasonal or if buyers were reaching their saturation level for digital music," the report says.

Apple rebutted Forrester's report, saying that iTunes sales continue to grow. But Apple did not offer specific numbers to counter Forrester's. Jobs' music manifesto certainly confirmed that consumers — people who already bought iPods, for Pete's sake — are simply not buying many iTunes songs.

It's not just an iTunes problem. In January, the International Federation of the Phonographic Industry — the global bureaucracy guarding music copyrights — said that online music sales in 2006 "nearly doubled." Which sounds amazing. Until you get to the part where the IFPI says that sales had tripled in 2005. So the growth rate had slowed.

"Downloads, as a business model for digital music, has failed," Dave Goldberg, VP of Yahoo Music, told a crowd at Digital Music Forum West late last year. "When you look at people who are buying downloads, it is older people who have money and time, and people who are doing it through gift cards."

How about subscription services, like Rhapsody and Napster? Not much joy there, either. Rhapsody reportedly has about 1 million subscribers. The rest, all together, have about another million. That's an audience share that approaches the 0.7 rating Animal Planet got when it aired Puppy Bowl III before the Super Bowl.

It's certainly not that people don't want to buy stuff on the Internet. Amazon.com's sales soared in 2006. Blue Nile is thriving selling diamond jewelry, and eBay sells millions of cars. Getting people to buy songs ought to be a snap.

And people want music. The Grammy Awards on Sunday were a giant celebration of music's popularity. People listen to more music in more ways than ever.

But for the majority of people, downloading songs is too hard and too frustrating. Some of that problem is the digital rights management (DRM) software that limits where and how songs can be played. It makes iTunes songs playable only on iPods, Rhapsody subscription songs playable only on certain devices, and so on.

The record companies believe DRM keeps people from pirating music, which may or may not be true. But DRM definitely keeps people from buying online music. As Jobs says, if consumers could buy music from any online store and play it on any device, the entire industry would thrive.

"The more you try to control music, the more you limit business opportunities," says Steve Waite, author of the book Quantum Investing and a professional musician.

Or as music artist Moby told me last week, "Personally, I see 2% of DRM as protecting copyright and 98% annoying consumers."

There are other reasons downloads are stalled. People who grew up with CDs — or vinyl LPs before that — like the packaging and cover art, and like to get songs deep in an album that are not hits but grow on you over time. At 99 cents a song, digital downloads don't offer enough value to give up the packaged CD niceties.

Especially when pirated music is so easily available for free.

If digital downloads are going to take off, they probably need to be DRM-free, simpler to buy and much cheaper. Then again, that will only happen with the record companies' blessing, and since they get 90% of their revenue from CD sales, maybe they just don't care about taking digital downloads to the mass market.

I guess we'll find that out if record label EMI, as rumored, decides to sell songs with no DRM.

After Jobs released his essay, I asked Apple spokeswoman Natalie Kerris why he chose to do it at that time. She said there was no particular reason. But Jobs never does anything for no particular reason. He manipulates the media and timing better than anyone in tech.

More likely, Jobs realized it was time to save iTunes.

Sunday, February 04, 2007

Second Life - The Next Frontier on the Web

Here's a nice piece from FORTUNE on Second Life, a virtual 3-D world that is emerging on the web. Click on link below for a glipse of the future.

http://money.cnn.com/2007/01/22/magazines/fortune/whatsnext_secondlife.fortune/