Maximum PC r&d RSS Feed

Intel Wants to put a Chip in Your Brain

Justin Kerr — Sun, 22 Nov 2009 15:25:23 -0600

Anyone who follows Intel closely knows that they don’t just pump out high end CPU’s, but they actually dedicate entire teams to “pie in the sky” ideas of what future technologies might look like. This could be anything from an x86 cluster of CPU’s to render video, or in this case, using your brain to control a computer. It may sound farfetched, but its something Intel and its researchers have been actively studying for sometime now.

Currently scientists are focusing on how the brain reacts when interacting with a computer, and then learning ways to interpret this data to execute commands on the machine. The idea here is to allow your thoughts to take over for your mouse and keyboard. Intel is of the belief that an implant would make this easier, though I’m not entirely sure how many volunteers they are going to get with that idea. “Eventually people may be willing to be more committed… to brain implants" said Intel’s Vice-President of future Technology, Andrew Chien. "Imagine being able to surf the Web with the power of your thoughts”

You may have your doubts, and so do we, but it might interest you to know that researchers have already made significant strides in the field of reading brain patterns, and have already identified certain words such as “bear” that cause everyones brain to react in a similar manner. “I think human beings are remarkably adaptive,” said Chien, “If you told people 20 years ago that they would be carrying computers all the time, they would have said I don’t want that, I don’t need that. Now you can’t get them to stop. There are a lot of things that have to be done first but I think implanting chips into human brains is well within the scope of possibility”. Chien speculates we will be lining up for implants as early as 2010.

Are you comfortable with this idea?

(Image Credit: howstuffworks.com)

White Paper: OLED Screens

Nathan Edwards — Mon, 02 Nov 2009 14:02:23 -0600

Organic light-emitting diode (OLED) screens offer better picture quality and draw less power than traditional LCDs. But what are OLEDs?

Organic light-emitting diodes, or OLEDs, are often touted as the next big thing in display technology, offering brighter colors, true black, lower power consumption, and better off-axis viewing than traditional LCD screens. They’ve popped up in gadgets from high-concept to mundane: The infamous Optimus Maximus keyboard, for example, utilizes many tiny OLED screens in its programmable and customizable keycaps, and both Sony’s new X-series Walkman and Microsoft’s new Zune HD have OLED screens. OLED technology has made great strides in the past 10 years, and cheaper and better manufacturing processes mean they’ve started appearing in everything from media players to phones to high-definition televisions—even keyboards. But what are OLEDs?

What’s Inside

In the simplest terms, LEDs (light-emitting diodes) emit light by running an electrical current through a diode. Diodes create unidirectional electric flow, moving electrons from the negatively charged cathode to the positively charged anode, creating electron holes, or spaces where electrons could be. Electrons flowing in drop into these holes and emit light. An organic light-emitting diode uses the same principle, but between the cathode and anode are two layers of organic semiconductor compounds: the emissive layer, near the cathode, and the conductive layer, near the anode (organic compounds are chemical compounds that contain carbon). The cathode sends (negatively charged) electrons into the emissive layer, while the anode draws electrons from the conductive layer, leaving positively charged “electron holes.” This creates a negatively charged emissive layer and a positively charged conductive layer, which attract each other, drawing electron holes to the emissive layer. The positive-charged holes and negative-charged electrons recombine, lowering the energy levels of the electrons, emitting light as a by-product. Simple, right?

From a development standpoint, OLEDs have a lot of potential. Organic chemistry is a fairly well-understood science—reds, blues, and greens were developed in a much shorter time frame in OLEDs than in regular LEDs. And new molecules that can be used in the layers, which have longer lifetimes and produce brighter colors, are being discovered frequently.

In an OLED, an electrical current causes electrons (-) to move from the cathode to the emissive layer, creating a negative charge in the emissive layer. The positively charged anode attracts electrons from the conductive layer, creating a positive charge in the conductive layer, which recombine with holes (+) in the conductive layer attract electrons from the emissive layer, which recombine with the electron holes, lowering the energy level of the electrons and emitting light as a by-product.

It turns out that OLEDs are great for use in displays, because the organic molecules that comprise the emissive and conductive layers can be deposited in very thin, large sheets onto a variety of substrates—from glass to metal to fiber—so that millions of individual OLEDs can be crammed together, row by row and column by column, into a very small space. Each of these OLEDs becomes one pixel of the display. The organic compounds can be deposited using several methods, depending on the type of organic molecule used in the display.

There are two types of OLEDs currently in production and development, differentiated by the size of the molecules in their organic compounds. Small-molecule OLEDs are usually manufactured via organic vapor phase deposition (OVPD)—the organic molecules are evaporated and carried via inert gas, then deposited on a substrate through a series of very small nozzles held near the substrate’s surface. Large-molecule, or polymer OLEDs, can be created via a process similar to inkjet printing—the polymers are dissolved into a solution and “printed” onto the substrate.

Advantages & Disadvantages

The advantages of OLEDs over traditional LCDs are many. First, unlike liquid-crystal displays, OLED pixels actually emit light, so they don’t require backlighting. Traditional LCD screens often utilize traditional LEDs or CCFLs for backlighting, which—in addition to increasing the thickness of the display to accommodate a light source—prevents the display from rendering true black, as even “black” LCD pixels are backlit. Since OLED pixels produce light when on and don’t produce light (or draw power) when off, a darker, richer black can be created. Having light-emitting pixels also enables richer colors, a broader color gamut, higher contrast, and a greater viewing angle than an LCD screen. Because “off” pixels don’t draw power, and because there’s no need for a separate light source, OLED displays require less energy to run. And because the organic molecules can be printed onto a variety of substrates, flexible displays are possible.

OLEDs, however, are not without their disadvantages. The manufacturing process is still expensive, so large OLED displays are rare—most OLEDs are used in small-screen applications, such as media players and smartphones, though HD displays up to 40-inches have been demonstrated. And the materials used in OLEDs don’t necessarily last as long as regular LCD displays—another reason they’re more frequently found on phones and media players, rather than computer monitors and televisions. Monitors are typically turned on for much longer stretches of time. And finally, the organic materials in OLEDs are extremely susceptible to water damage, so displays must be well-sealed.

OLED to the Future

What’s next for OLED technology? The European Union, among others, is investigating the use of OLEDs as cheap solid-state lighting to replace incandescent bulbs. Their stated goal is to create a 100x100cm square of OLED material that creates 100 lumens per watt of power, has a working lifespan of at least 100,000 hours, and costs less than 100 euros per square meter to produce.

OLEDs have found their way into concept cars, lighting fixtures, PMPs, and laptop prototypes, with the latter expected to enter production by Q3 2010. As manufacturing processes become less expensive, OLED displays could start to replace LCDs, not just in media players and phones, but also in notebook computers, monitors, and televisions, on a much larger scale.

Redmond Shows Off Cure for Teleconferencing Headaches at TechFest 2009

Mark Edward Soper — Wed, 25 Feb 2009 21:55:15 -0600

If you've ever been subjected to a babel of echoing voices during a teleconference, Microsoft Research is working on a solution. As demonstrated (link requires Microsoft Silverlight) at this week's TechFest, MR's audio spatialization project enables a PC with stereo speakers to spatially separate different members of a teleconference. Audio spatialization's been used for years in 3D gaming, but Microsoft Research has added a new twist: to make it work for teleconferencing, it's also added echo cancellation. As researcher Zhengyou Zhang puts it:

Audio spatialization uses speakers to create the illusion that call attendees have different locations spatially. This allows you to use the audio sense you already have, that you normally use in conversation, to isolate who you’re talking to, and to associate a location in space with a particular individual... In a conference where there are multiple voices coming out of multiple speakers, it becomes important to eliminate the echoes that might naturally occur.

See it for yourself, then hit comment and sound off.

Microsoft Research Shows Off Latest Projects at TechFest 2009

Mark Edward Soper — Tue, 24 Feb 2009 22:01:02 -0600

Microsoft Research's latest chance to shine is this week's TechFest 2009. Microsoft Research has a long list of innovations, including the Microsoft Surface touch-sensitive interface, the Unwrap Mosaic video editor, the Songsmith music composing utility, Image Composite Editor, and many more. TechFest serves two purposes: it makes sure that everyone at Microsoft can tap into what's being developed at Microsoft Research, and it acts as a sort of high-tech equivalent to an auto show, demonstrating the concepts that might (or might not) make their way into future products from Redmond.

This year's TechFest features projects as varied as combining multiple cell phone videos to create a high-res version; using digitized books on video DVD to create a high-capacity, low-cost library and school resource for developing countries, and ways to create Augmented Reality, which overlays digital data with real-world information, to name just a few.

So, how important are Microsoft Research projects to Microsoft's future? As Microsoft Research head Rick Rashid sees it, the investment Microsoft makes in research is "really about an investment in survival." What do you think is the coolest concept at this year's TechFest? Hit Comment and tell us.

Terabit Ethernet, Concepts Proven

Justin Kerr — Sun, 15 Feb 2009 11:44:15 -0600

Gigabit Ethernet may still outrun all but the most extreme SSD Raid configurations, but researchers can never rest on their laurels. Always hoping to invent the next big thing, scientists now have their sights set on Terabit Ethernet to help quell our insatiable hunger for bandwidth. A team from Australia, Denmark, and China has combined their efforts to demonstrate terabit-per-second speeds using fiber optic cables, laser light, and an unusual material named chalcogenide.

The group documented the results of its most recent trial in a white paper published in the February 16th 2009 issue of Optics Express. Though the technology is promising, Ben Eggleton, research director for CUDOS (Center for Ultrahigh bandwidth Devices for Optical Systems), points out the current limitations. “The problem isn't injecting that much high speed data into an optical strand, called multiplexing, but retrieving data at such high rates”. Conventional electronics are capable of injecting dozens of 10 Gbps streams, but trying to retrieve these streams any faster than 40 Gbps is beyond our current capabilities.

The breakthrough here however isn’t in the speed itself, but in proving the concept. Until the processing hardware catches up with our transmission capabilities, you won’t be finding this in routers anytime soon. Eggleton speculates that these concepts can be adapted to achieve slower and more manageable results, but the goal of this experiment was simply to prove that it was possible using fully photonic chips built using the same methods employed by current CMOS circuits. "It's years to complete," Eggleton said, taking these research efforts into a production technology. But these demonstrations "are starting to establish this is a serious proposition."

White Paper: Virtual Machines

Michael Brown — Thu, 01 Jan 2009 14:00:00 -0600

System virtual machines fall into two categories: Type 1 hypervisors (left), which run directly on the host hardware, and Type 2 hypervisors (right), which run on top of another operating system. Both are capable of running multiple independent instances of one operating system or different operating systems, all of which behave as though they are solely in control of the system.

Can a computer exist without hardware? It can if it’s a virtual machine. A virtual machine is software that’s capable of executing programs as if it were a physical machine—it’s a computer within a computer. Virtual machines can be divided into two broad categories: process virtual machines and system virtual machines.

A process virtual machine is limited to running a single program. A system virtual machine, on the other hand, enables one computer to behave like two or more computers by sharing the host hardware’s resources. A system virtual machine consists entirely of software, but an operating system and the applications running on that OS see a CPU, memory, storage, a network interface card, and all the other components that would exist in a physical computer. For the remainder of this discussion, we’ll use the term “virtual machine” to refer to a system virtual machine.

Software running on a virtual machine is limited to the resources and abstract hardware that the virtual machine provides. Since a virtual machine can provide a complete instruction set architecture (ISA, a definition of all the data types, registers, address modes, external input/output, and other programming elements that a given collection of hardware is capable of working with), a virtual machine can simulate hardware that might not even exist in the physical world.

Using virtual machines, a computer can run several iterations of an operating system—or even several different operating systems—with each OS isolated from and oblivious to the existence of the others. The only requirement is that each operating system must be capable of supporting the underlying hardware. And, of course, there must be enough resources (memory, hard disk space, CPU cycles, and so on) to support everything. You could use a virtual machine to run Linux on top of Windows, for instance, or you could run two versions of Windows and use one as a sandbox for testing software you wouldn’t trust on a “real”

More Powerful than a Supervisor

The software that manages this trick is known as a hypervisor. A Type 1 (native) hypervisor is a program that runs directly on the host hardware, i.e., as an operating system in and of itself. Microsoft’s Hyper-V, formerly known as Windows Server Virtualization, is one example of a Type 1 hypervisor. A Type 2 (hosted) hypervisor, such as Microsoft’s Virtual PC 2007, runs on top of another operating system.

IBM developed the technology for its big-iron mainframe computers in 1967, but the Intel x86 architecture at the foundation of IBM PC-compatible machines was not well suited for running hypervisors. Achieving full virtualization required exceedingly complex code, which hampered runtime performance. Although it remained a fixture in mainframe and midrange computer systems, virtual machine technology saw very little progress during the 1980s and 1990s.

In the last few years, however, AMD and Intel both developed extensions to their x86 architectures that render newer CPUs much more suitable for running hypervisors. AMD has dubbed its extensions AMD Virtualization (AMD-V); Intel calls its extensions Intel Virtualization Technology (Intel VT). AMD-V is present in many newer AMD CPUs, including the Athlon 64 and Athlon 64 X2 (socket AM2 only), the Phenom X3 and X4, and second-generation Opteron server parts.

You’ll find Intel VT in about half of Intel’s Core 2 Duo desktop processors (the E6600 through E6850, and the E8200 through E8600), all of its Core 2 Quad and Core 2 Extreme desktop processors, and its quad-core Xeon and Itanium server procs (the Itanium version is formally known as Virtualization Technology for IA-64). Intel’s upcoming Core CPU will feature Intel’s VT-d (Virtualization Technology for Directed I/O), which will enable guest virtual machines to directly use peripheral devices, such as a network interface device. Although AMD-V and Intel VT are similar, they’re not compatible; a hypervisor that supports only AMD-V will not take advantage of the virtualization extensions in an Intel CPU and vice versa. Fortunately, hypervisors that support both sets of extensions are common.

Applications for Virtual Machines

What are virtual machines good for? The most common application today is server deployment. A virtual machine can make much more efficient use of a server’s hardware by running several instances of the same operating system and the same applications in parallel, or even different operating systems and applications.

In either scenario, each instance thinks it has sole access to the hardware and behaves accordingly. The hypervisor dynamically assigns virtual resources (such as processors and memory) to physical resources so that the hardware is never left idle. Virtual machines are also useful as test platforms: System designers and application developers can experiment with new code without disrupting or interfering with the usual production environment.

But virtual machines are useful for individual users, too. Experimenting with different operating systems—such as Linux—on one computer is just one example. Trying out new software—especially shareware—is another. If a program renders your system unstable, you can blow away the virtual machine without any consequences. Or if you’re paranoid about privacy, you could create a virtual machine explicitly for web browsing: Isolate all your personal information on one installation that you never use for web surfing. Fire up the virtual machine when you do want to browse the web and tracking cookies, spyware, and any other Internet detritus you encounter will be trapped there, where it can’t harm your production environment.

Getting started with virtual machines is certainly cheap enough: Several programs are available for free, including Microsoft’s Virtual PC and Sun Microsystems’s VirtualBox (the latter of which is capable of running Windows as a guest operating system running on Linux).

Everything You Need to Know About Photovoltaic Cells

Michael Brown — Mon, 22 Sep 2008 13:49:11 -0500

In one second, the nuclear fusion process taking place inside the sun produces enough energy to satisfy the needs of the earth’s population for nearly 500,000 years. Photovoltaic cells are capable of capturing some of that energy and converting it into usable electricity; unfortunately, today’s technology can’t do this very efficiently.

French physicist Edmond Becquerel first described the photovoltaic effect in 1839. He discovered that some materials were capable of producing small amounts of electricity when exposed to sunlight. The first photovoltaic cell, however, wasn’t created until 1883, and more than 70 years passed before the next major scientific advance took place, when researchers at Bell Labs developed the first crystalline silicon photovoltaic cell in 1954.

Most modern photovoltaic cells are still manufactured from silicon, the same semiconductor material used to produce GPUs, CPUs, and other integrated circuits. The majority of commercial photovoltaic cells are manufactured from crystalline silicon—either single- or poly-crystal silicon. The latter are less efficient than the former, but their lower manufacturing cost largely makes up for the conversion shortfall.

The bulk of the progress that’s been made since the 1950s stems from the efficiency at which absorbed light is converted into electricity. The Bell Labs product was capable of just 4 percent efficiency; today’s commercial products are approaching 20 percent efficiency.

Photons from the sun pass through the cell’s n-type layer to strike atoms in the p-type layer,
dislodging some of those atoms’ electrons in the process. The freed electrons move up toward the n-layer,
creating an electrical current that can be stored or service an electrical device.

The Photovoltaic Process

A photovoltaic cell is created by sandwiching two silicon wafers: an n-type layer and a p-type layer. The n-type layer exhibits a negative electrical charge and has an excess of electrons, while the p-type layer exhibits a positive electrical charge and has a shortage of electrons. The two layers are separated by an n-p junction. The cell is then attached to a backplane, a layer of metal used to physically reinforce the cell and provide an electrical contact on its bottom. A second electrical contact is placed on the top of the cell to create an electrical circuit. The cell is then treated with an anti-reflective coating to compensate for silicon’s otherwise shiny nature.

As photons—particles of light—hit the photovoltaic cell, they pass through the n-type layer and strike the p-type layer, where they are either absorbed by the silicon atoms, reflected, or pass straight through the material. Absorbed photons knock electrons loose from the silicon atoms, leaving empty “holes,” which are filled by electrons further back in the circuit. The loose electrons flow through the electrical contacts on the p-type layer to the contacts on the n-type layer. This flow of electrons produces an electric current that can be drawn off and stored in a battery or used to power an electrical device.

An array of cells is electrically connected and mounted into a frame to form a photovoltaic module. A narrow metal grid is applied to the top of the module to transport electrical energy, and a sheet of glass or plastic is placed on top to protect the cells from the environment (everything from bad weather to bird droppings and stray baseballs). A group of interconnected modules is known as an array.

Photons contain varying amounts of energy, depending on their wavelength. Within the visible spectrum, red light possesses the least amount of energy while violet light has the most. The same goes for the invisible spectrum: Infrared light possesses very little energy but ultraviolet light contains a great deal of it.

Most modern photovoltaic cells are capable of converting only high-energy photons into electrical current, which explains why mainstream solar panels are so inefficient. One of the most promising ideas for increasing the efficiency of solar energy is to stack cells with different properties on top of one another. This way, high-energy photons can be captured by a cell on the top of the stack, while lower-energy photons pass through to subsequent cells that are better suited to those photons’ wavelengths.

AC/DC

The electrical devices in your home (appliances, computers, air conditioners, lights, and so on) operate on alternating current (AC), but a solar array produces direct current (DC). The solution is to install an inverter that converts the solar array’s DC into AC. Inverters are designed to power off when there isn’t enough electrical current for them to operate, e.g., at night.

Solar panels produce the most power in the presence of direct sunlight, but they’ll produce some energy on cloudy or even rainy days. They can’t produce any juice at night, of course, so you’ll need some means of storing the electricity that they create when the sun is shining. Batteries can provide total independence from your local electric company, enabling you to potentially live “off the grid,” but this solution presents a host of environmental problems, and there’s no guarantee it will provide all the energy you’ll need. The more practical alternative is to tie your system into the electrical grid.

In a grid-tied system, you sell the excess energy your solar array generates to your local utility, and you buy back the electrical power you need for your home. With this method, the utility acts like an unlimited energy-storage system, giving you all the power you need whenever you need it. The inverter is connected to the meter the electric utility uses to measure your consumption, which means your meter will spin backward whenever you generate more than you consume.

For most households, the reward for going solar is more feel-good than financial: It could take a decade or longer to recoup the investment in even a moderate-size system. That situation is changing rapidly as the escalating cost of producing electricity from fossil fuels moves in inverse proportion to the cost of deriving energy from the sun.

Everything You Need to Know About GDDR Memory

Michael Brown — Tue, 16 Sep 2008 01:11:36 -0500

We invariably refer to the video memory in modern videocards as GDDR, differentiating it only by version (GDDR2, GDDR3, GDDR4, and now GDDR5), but the technology’s full acronym is actually GDDR SDRAM, which stands for Graphics Double Data Rate Synchronous Dynamic Random Access Memory.

“Double data rate” describes the memory’s capacity for double-pumping data: Transfers occur on both the rising and falling edges of the clock signal. This endows memory clocked at 800MHz with an effective data-transfer rate of 1.6GHz. “Synchronous” refers to the memory’s ability to operate in time with the computer’s system bus. This allows the memory to accept a new instruction without having to wait for a previous instruction to be processed, a practice known as instruction pipelining.

GDDR2 memory was never a very popular solution among GPU manufacturers: The technology required 2.5 volts to power its input buffers and core logic (i.e., VDD voltage), which is the same as GDDR. GDDR2 operated at much higher clock speeds than its predecessor, however, which produced a tremendous amount of heat. The fact that GDDR2’s VDDQ voltage requirement (the electricity needed to power the memory’s output buffers) was only 1.8 volts didn’t compensate for this problem.

Survival of the Fittest

GDDR3—an open standard developed by ATI in conjunction with the standards organization JEDEC Solid State Technology Association—is the most widely used graphics memory technology in use today. Ironically, Nvidia introduced the first graphics processors designed to use GDDR3: The GeForce FX 5700 Ultra, followed by the GeForce 6800 Ultra. ATI didn’t deploy a GDDR3 solution until it shipped the Radeon X800.

GDDR3 improved on previous GDDR designs by supporting higher clock speeds while requiring less power. These chips consume less electricity, so they produce less heat and can rely on simpler cooling hardware (GDDR3’s VDD and VDDQ voltage requirements are both 1.8 volts). GDDR3 also has separate read and write data strobes, which contributes to a much faster read-to-write ratio (meaning the turnaround from a read operation to a write operation occurs much more quickly) than GDDR2 supported. GDDR3 chips have a hardware reset feature that can wipe their memory clean to start receiving new data should such an operation be necessary.

ATI and Nvidia (in conjunction with JDEC) both had a hand in establishing the specification for the next generation of graphics memory, GDDR4, but Nvidia has so far decided not to use the new technology in any of its reference designs. ATI, meanwhile, incorporated the new memory first in its Radeon X1950 XTX cards and subsequently in several models of its Radeon HD 2000, 3000, and 4000 series.

Evolutionary Dead End

GDDR4’s improvements over GDDR3 were mostly incremental. It seemed to offer a power advantage in that it could operate with just 1.5 volts, compared to GDDR3’s 1.8 volts. Board designers, however, quickly discovered that they needed 1.8 volts anyway to ensure stability at higher clock rates.

Two other GDDR4 enhancements are more significant in that they increase the memory’s overall performance: The new memory doubled the size of GDDR3’s prefetch scheme from 4 bits to 8 bits, and its burst length was locked at 8 bits (GDDR3 supports either 4- or 8-bit burst lengths). Prefetch enables the memory chip to anticipate the need for data and grab it before the GPU asks for it, reducing the time the processor has to wait. Burst length defines the amount of data sent in burst mode, a process in which data is transmitted without waiting for input from another device, such as the GPU.

GDDR4’s 8-bit burst length might be one reason Nvidia ultimately passed on this type of memory: Nvidia’s processors support only 4-bit burst lengths. With ATI (now AMD) being the only major customer for GDDR4, just two manufacturers—Samsung and Hynix—decided to manufacture it. This circumstance has kept the price of the memory relatively high.

Successful Mutation?

GDDR5 is the next major development in graphics, and as with GDDR4, AMD’s ATI division has already paired it with its higher-end GPU: the Radeon HD 4870. Nvidia continues to hang back, professing satisfaction with the performance of GDDR3.

GDDR5 requires just 1.5 volts of electrical power, which should make the memory run cooler—a feature that could aid in overclocking, reduce manufacturing costs, and extend battery life if used in a notebook PC. The new memory’s prefetch and burst length remain the same as that of GDDR4: 8 bits on both counts.

GDDR5 technology supports densities ranging from 512Mb to 2Gb, so it would require just four 2Gb modules to create a 1GB frame buffer (here again, however, real-world parts are currently limited to 512Mb and 1Gb). Boasting a raw theoretical data rate ranging from 3.6Gb/s to 6Gb/s (although we won’t see that upper limit for several years), GDDR5 promises to deliver twice the memory bandwidth of GDDR3 running at the same clock frequency.

More practically, that high data rate also enables a GPU manufacturer to achieve nearly the same memory bandwidth with an economical 256-bit interface as it would by building a much more expensive 512-bit bus into its GPU.

Nvidia’s professed ambivalence toward GDDR5 hasn’t stopped a third major memory manufacturer—Qimonda—from joining Hynix and Samsung in the market for GDDR5 memory. Hmm, is anyone taking bets that Nvidia’s next-generation GPU will tap GDDR5?