Soft Machines claims its cutting-edge VISC CPU cores can outperform Intel, ARM in performance per watt

More than a year ago, we coveredSoft Machines VISC (Variable Instruction Set Computing) and the company’s long-term goal to improve efficiency. VISC’s argument is that by creating a middleware software layer that can translate single-threaded code into parallel workloads that are executed by multiple virtual cores, it can improve overall execution efficiency and reduce power consumption. Or at least, that’s been the claim.

Soft Machines has now revealed more performance data on how it expects its first VISC core, Shasta, to perform, as well as information on the upcoming Shasta+ and Tahoe CPUs.

The first Shasta core will be available this year, with 1-2 virtual cores on a dual-core configuration, or an SMP block of 2-4 VCs with a quad-core configuration. The CPU has a 64-bit ISA and should be clocked at 2GHz. By 2017 Shasta+ will move to 10nm with support for more virtual core instances, followed by a new architecture, Tahoe, in 2016.

This graph captures much of what Soft Machines believes makes its hardware appealing. The company is basically arguing that by virtualizing CPU resources and breaking even single-threaded workloads into pieces that can be spread to different cores (with hypothetically different resources and capabilities) it can realize greater efficiencies than CPU architectures that rely on dynamic frequency and voltage scaling (DFVS).

The big question to answer, I think, is how much of an overhead penalty SoftMachines pays for its virtualization, and what kinds of workloads it can effectively execute on its cores. SPEC is a decent cross-platform benchmark, but it’s also susceptible to hand-tuning and careful optimization. SoftMachines’ documentation states that the same GCC 4.9 settings were used for all processors, but SPEC runs aren’t the same as commercial software deployments.

Now the Shasta results being shown here are simulated, but again, SoftMachines claims to be using the same model they adopted for simulating the performance of their proof-of-concept 28nm core. The simulation method proved accurate for that chip, within 5% on performance and 10% on power. In theory, therefore, the Shasta, Shasta+, and Tahoe results should match as well.

We see plenty of CPU announcements come and go in the journalism business, but Soft Machines has been flying largely under the radar since 2014. They’ve made a few additional announcements, but most of the company’s efforts have apparently been on improving its products as opposed to its media profile. I’m genuinely curious to see if their virtualization approach can actually yield benefits in real-world scenarios, particularly given the difficulty that companies like Intel have had with increasing overall performance. Breaking workloads up dynamically and executing them across virtual “cores” could be more power-efficient than scaling single cores up and down by clock speed, but demonstrating that efficiency in real-world tests will still take some additional work.

Since Soft Machines doesn’t build its own CPUs or SoCs, we’ll have to wait for partner silicon to come to market before we can draw firmer conclusions about whether this approach can improve performance.

MIT and Texas Instruments develop hack-proof RFID chip

More and more devices are showing up with RFID chips built-in, but there is concern that the data on these chips could easily stolen. After all, an attacker doesn’t even need to physically have possession of the RFID chip to get information from it. Protecting the data on the chip with a secret key will thwart a casual data thief, but there are still ways around that. Researchers from MIT have developed a new type of RFID chip (manufactured by Texas Instruments) that they claim cannot be hacked by any current means. They manage this with a combination of integrated power and data storage not previously seen in RFID technology.

Most RFID hacks are based on what is known as a side-channel attack. Basically, by analyzing the pattern of power usage and memory utilization, it’s possible to extract the cryptographic key from a system. Side-channel attacks only leak a little data for each repetition of an algorithm, so you need to run the attack many times to get a full key. One way to thwart these attacks is to rotate the private key frequently, but a determined hacker can get around this with a so-called power glitch attack, and that’s what the RFID chip from MIT is designed to block.

Power-glitch attacks involve cutting power to a device right before it can rotate its secret key. That allows the attacker to run the same side-channel attack numerous times to get the key. A power-glitch attack can be used on various devices, but RFID chips are particularly vulnerable as they don’t have a built-in power source. Instead, they’re powered by induction from the reader. The highly secure RFID developed by graduate student Chiraag Juvekar and his faculty advisers has an integrated power supply and non-volatile memory to guard against this exact scenario.

This chip takes advantage of a material called ferroelectric crystals. They consist of molecules arranged into a lattice where positive and negative charges naturally separate. Applying an electric field can flip the charges to one direction or the other, thus representing a bit of information. A ferroelectric crystal can also operate as a capacitor for storing power; this is the voltage difference between the lattice’s negative and positive poles.

Texas Instruments’ manufacturing process can create banks of 1.5v and 3.3v cells on the RFID composed of ferroelectric crystals. When a power glitch attack is attempted on this chip, the 3.3v cells act as an energy source allowing the chip to store the data it’s working on in the 1.5v cells. When power is restored, the first thing the chip does is recharge the 3.3v cells in case power is lost again, then it picks up where it left off with the saved data. If it’s trying to rotate the secret key, it continues doing so and makes the attack useless.

The team speculates that this technology, if adopted widely, could make RFID chips considerably more secure. The storage and power requirements increase cost, and the output rate is a bit slower than conventional chips. However, the team found that it could still produce 30 readouts per second, which should be fine for most RFID applications.

The coating conducts electricity on a contact lens

Google Glass was somewhat beset with controversies -- often thought to be unsightly, the tech was also banned in cinemas,declared "no safer than texting"when worn by drivers, andsubject to privacy debates. They may be resigned to the past, however, as a new contact lens is developed that can act as a computer screen. 

The proof of concept, published in Applied Materials and Interfaces and designed by researchers at the University of South Australia, details a polymer film coating that conducts electricity on a contact lens. This could, the researchers say, eventually allow the development of a tiny electrical circuit that could sit on a lens worn by a person, allowing someone to read text or project a computer screen direct from their eye.

Referring to the technology as a "game changer", Professor Drew Evans, who worked on the project, said the method would be a safe way of bringing wearable tech even closer.

"We're talking about anything from a simple sensor that can measure the amount of glucose in the blood through to actually creating electronic displays," he said. "So rather than having something like a pair of glasses that acts like a computer, you can generate images directly onto your contact lens". 

The contact lens works by coating a normal lens with a thin, "biocompatibilising" film, which allows a conductive electronic circuit to work with the body. The lens is pretreated with plasma in order to facilitate both the attachment of the film and its adherence to the eye. The polymers used in the lens have also been used to create 'smart windows', which darken and lighten and may have application as camouflage for the military.

The research has taken several years, and the team now hope that the lens will enter further testing and potentially be available to consumers in the UK and elsewhere. Their next step, they say, is developing complementary technology that can read and translate the information transmitted by the conducting polymers. 

"What is really significant is that the materials we are developing are not only safe but also have the potential for a range of personalised health monitoring applications that could make life simpler for people struggling with chronic health problems," Evans said.

Optical magnetometer performs first-ever noninvasive detection of nerve impulses

It’s surprisingly difficult to pinpoint what a given nerve is doing at any given moment. The electrochemical dance of neuronal function never stops, and it’s synced to the beat of the default mode network more strongly than to any external clock. Neurons are constantly receiving and transmitting information, but an electrical action potential is measured in millivolts and only changes the magnetic field around a neuron by a few picoTesla.

Up ’til now, this tiny variance has meant we needed invasive methods to get any information whatsoever about the function of living nerves. Now, scientists from Denmark have used a game-changing device for an all-new neuronal imaging technique: an optical magnetometer that can measure a single nerve’s function from outside the body — with quantum-level precision.

Today’s best methods of discerning a nerve’s function are still pretty invasive. Either we have to stick an electrode into a nerve, which is pretty disruptive, or we have to actually dissect a creature and thread one of its neurons through a tiny conductive coil on a machine called a SQUID. Either way, we’re not very good at getting information out of nerves in vivo. And neuroscience has needed another way of looking at nerves in real-time from outside the body, preferably at room temperature.

That’s where the optical magnetometer comes in. These devices work because they use a laser that detects the change in polarization of gaseous cesium atoms when they’re subjected to a changing magnetic field. The flux induced by an electric action potential causes a flutter in the polarized light, which the magnetometer can detect. And the sensitivity of these devices is unparalleled: Their resolution is limited only by quantum effects like the quantum shot noise of light.

The breakthrough here is in the application. This kind of magnetometry has never before been applied to living cells, in part because the magnetic flux generated by an action potential is so very small: nine orders of magnitude smaller than your average fridge magnet. That kind of precision is pretty hard to get at all, to say nothing of using it in vivo. But the combination of how it uses the laser and the tiny size of the sensor end means this device can point at a nerve and see what that particular nerve is doing, to the exclusion of fibers around it. Jensen and co. tried it out on a frog, and used the magnetometer to detect when its sciatic nerve was firing.

A discovery like this has the potential to change the entire brain-imaging field. The team that did this project notes “The magnetometer [is] perfect for medical diagnostics in physiological/clinical areas such as cardiography of fetuses, synaptic responses in the retina, and magnetoencephalography,” or presumably anything else that requires noninvasive brain imaging in the time domain. And it won’t be long until we’re using this technique on humans. This is a clear step forward for both basic research and the biomedical sciences alike.

Google taps chipmaker Movidius to include machine learning for phones

A main reason electronic devices around us are getting smarter, and hopefully more useful, is machine learning. By building elaborate models of data, then training those models, tasks as unrelated as facial recognition, language translation, and autonomous driving can be down successfully. In addition to the need for huge amounts of computing power to train those systems, they all tend to be both very power and processor intensive to run. That has kept most of them tethered to plug-in devices — like the Kinect — or requiring large batteries, like those found in a car. For example, Nvidia’s Drive PX 2 trunk-mountable car computer will require liquid cooling. For mobile devices that has meant a constant connection to the cloud, with raw data sent up, analyzed at the data center, and the results returned.

Google has been trying to change this dynamic with Project Tango, a mobile device that can do real-time mapping and some object tracking, while running off only a small battery. To accomplish that, it tapped a new kind of processor, the Video Processing Unit (VPU) chip Myriad 1 from startup Movidius. By moving the processor-intensive tasks associated with computer vision into a specially designed chip, Myriad increased the performance of, and decreased the power requirements for, the vision-related functions of the Tango device. Movidius claims at least a factor of 10 savings in power, along with an 80% reduction in both space and cost over competing technologies — all compelling stats when it comes to mobile device design.

Beyond Project Tango: Using Movidius for mobile machine intelligence

Now, Google has broadened its relationship with Movidius, announcing that it will be using the company’s newest and most powerful VPU, the Myriad M2450, to help bring more intelligence to a wider array of mobile devices. The Myriad isn’t limited to running vision-related applications, either. Google will use Movidius’s software development environment to port its advanced neural computation engine to the chip, so that a wide-variety of deep-learning-based algorithms can be run in real time.

Being able to run deep-learning-enabled tasks locally will reduce dependence on the cloud, thus reducing latency and privacy issues. For example, your phone could recognize your friends in a photograph without you needing to upload it to the cloud. Remi El-Ouazzane, Movidius CEO, explains, “The challenge in embedding this technology into consumer devices boils down to the need for extreme power efficiency, and this is where a deep synthesis between the underlying hardware architecture and the neural compute comes in.”

Unfortunately, there aren’t any details yet on any new Google products that will use the Movidius chips (and there was no mention of them at the Lenovo and Google Project Tango phone announcement), but given the importance of computer vision and machine learning to the future of mobile devices.

Graphene acting as‘optical capacitors’ enable chips to integrate biophysics and semiconductors

Neuromorphic chips are the new breakthrough in semiconductor research. These chips consist of networks of transistors that interact the way neurons do, enabling them process analog input, like visual information, faster and more accurately than silicon chips can.

We’ve been inching our way closer to neuromorphic circuit architecture for years, precisely because it’s so much faster than current von-Neumann-architecture silicon at critical stuff like image processing. The overlap between neuronal and silicon processing is significant — it’s easy to sketch out a metaphor between nerve fibers and bundled wires, myelin and insulation, synapses and logic gates.

Optoelectronics extends the metaphor even further, comparing photons moving through a laser transistor with neurotransmitters crossing the synaptic gap. Now researchers from Princeton have demonstrated a way of using graphene as an optical capacitor, to stabilize the function of laser transistors in optical neuromorphic circuits.

There are certain key differences that we’ll have to solve before we can make a processor that works enough like a brain to get the advantages. For example, neurons transmit information electrically through action potentials, or “spikes.” Spikes are all-or-nothing, very binary, so they have to encode information in the time domain. But the firing rate of a neuron isn’t limited by a central clock cycle: Instead, the firing frequency of a neuron encodes the intensity of the signal it’s sending. And neurons are an analog system, so they’re fast. But von Neumann chips are clock-limited, and we can’t go on cramming in transistors forever. We have to find other ways to make computing faster.

The answer, according to the authors of this recent study, is to stuff graphene into a laser. This allows the graphene to trap photons, turning it into an optical capacitor. When boosted in this fashion, the laser can “spike” at picosecond timescales. The IEEE notes that “Graphene, it turns out, makes a good saturable absorber because it can take up and release a lot of photons extremely fast, and it works at any wavelength; so lasers emitting different colors could be used simultaneously, without interfering with each other—speeding processing.”

Using graphene “sponges” in this fashion makes the lasers more optoelectronically accurate, and could be used to output multiple photons of different wavelength at the same time, without any interference.

At the end of Moore’s law, processors designed to mimic neurons and neural circuits could offer huge benefits in terms of power consumption and scalability. Optoelectronics — think fiber-optics and laser transistors — are a promising way of going about this, because photons are faster than electricity moving along a trace. Combine optoelectronics with a neuromorphic chip architecture, as these graphene capacitors demonstrated in the latest Nature Scientific Reports, and we may have a triple threat: a laser transistor that’s fast enough to handle input from analog sensory arrays in a neuromorphic circuit.

Shocking! 'Electric Eel' Fibers Could Power Wearable Tech

Stretchy fibers behaves like electric eels could be woven into clothing to power wearable technology one day, new research reveals.

Experimentally, these flexible fibers produces sufficient power to run electronic lights and watches.

The new fiber is interesting because it borrows a leaf from nature to "solve real-world problems and even surpass nature in some aspects," said study lead author Hao Sun, a materials scientistat Fudan University in Shanghai. [Top 10 Inventions That Changed the World]

High voltage
Electric eels (Electrophorus electricus) are able to generate deadly shocks to stun prey and defend against predators. These fish have cells known as electrocytes, which store and release electrically charged ions to generate powerful electric fields.

By themselves, electrocytes in electric eels generate low voltages of only about 0.15 volts. However, in eels, thousands of these disclike electrocytes line up, working in concert to produce deadly shocks of up to 600 volts, or about five times the voltage emitted from a U.S. electrical outlet.

Sun and his colleagues wanted to harness the power of the electric eel in a man-made material. To do so, they created fibers that resemble the shocking creatures' ability to stack up tiny voltage-producing cells in concert.

These fibers are capacitors, meaning they alternate pairs of electrical conductors and electrical insulators, or materials that block the flow of electricity. Capacitors store electric charge on the surfaces of the conductors, and can capture and release energy much more quickly than batteries can, although they usually store less energy than batteries do.

The scientists fabricated the capacitors by first wrapping sheets of carbon nanotubes around elastic rubber fibers 500 microns wide, or about five times the average width of a human hair. Carbon nanotubes are pipes only nanometers, or billionths of a meter, in diameter that possess remarkable electrical and mechanical properties.

The researchers made sure that the electrically conductive carbon nanotube sheets did not completely cover the electrically insulating rubber. Instead, there were gaps where the insulating rubber was exposed. Such gaps are key, because capacitors consist of both conductive and insulating units.

Then, the scientists applied patches of electrically conductive electrolyte gel onto these fibers. The pattern of patches the researchers used converted the fibers into capacitors.

The more alternating segments of electrically conductive nanotube sheets and electrically insulating rubber gaps a fiber had, the greater the voltage it could generate. A fiber about 39 feet (12 meters) long could generate 1,000 volts, the researchers reported online Jan. 14 in the journal Advanced Materials.

Previous research also sought to mimic electric eels by connecting many electrocyte like units together. However, those units were impractical because they were strung together with metal wires, and generally had poor flexibility, the researchers said. This new device instead connected all of its electrocyte like units together on a single fiber.

"We thought these findings provide an efficient strategy for the advancement of flexible electronics and wearable devices," Sun told Live Science.

Power fiber

The elastic fibers could stretch up to 70 percent more than their usual length without losing their electrical or structural properties, the researchers said. The team also showed that the fibers could be woven together with conventional elastic fibers to create fabric that could be incorporated into clothes.

The researchers suggested that the eely fibers could help power miniature electronic devices. For example, in experiments, they created energy wristbands to power electronic watches, and wove fibers into T-shirts to power 57 light-emitting diodes (LEDs). In the future, these energy fibers "might be incorporated into our daily clothes to power our wearable devices, such as the Apple Watch and Google Glass," Sun said.

The scientists also connected their capacitor fibers to fiber-shaped solar cells to create material that could both harvest and store energy. In experiments, these combinationfibers generated 10 volts of electricity when exposed to light — enough to power some types of small electronic devices, they said. Solar cell fibers could also recharge battery fibers in wearable devices, the researchers said.

Air umbrella produces a "force field" of air to keep you dry

People certainly haven't been afraid to try and reinvent the umbrella over the years. There was the solar-powered Booster Brolly, the windproof Rainshaderand the lopsided Rain Shield, just to name a few. But now a team of Chinese designers is looking to do away with the awkward metal poles and canopy entirely, relying instead on a "force field" of air to keep you nice and dry.

The idea of using air flow to shield users from pelting rain has been explored before. A pair of designers developed a similar concept back in 2010, though this version seemingly never made it to market.

The latest take on the Air umbrella uses a lithium battery to power a motor and fan, which creates a cycle of air flowing through its tip to deflect the drops and form the protective layer. So how much protection are we talking about exactly? Well, the designers say at the current stage of development the umbrella will provide enough shelter for two people, though if the rain isn't too heavy more people can be squeezed underneath.

The team is working on three models, all resembling something of a detached shower head. The first claims to be designed for females and is 30 cm (1 ft) in length and 500 g (1.1 lb) with a battery life of 15 minutes. The second is 50 cm (1.64 ft) long, weighs 800 g (1.76 lb) and has a 30 minute battery life. The third is extendable, ranging from 50 cm (1.64 ft) to 80 cm (2.62 ft), weighing 850 g (1.87 lb), also with a 30 minute battery life.

With these battery times you wouldn't want to be stranded in a rainstorm, but they might just be enough to get you home from the bus stop without becoming a cold, shivering mess.

The team is looking to raise US$10,000 on Kickstarter for the production of its Air umbrella and looks to be having reasonable success to date. At the time of writing it has attracted almost $12,000 in pledges with 10 days to run on the campaign. Most of the early bird pledges are already gone, with $108 the minimum pledge to reserve the shorter "female" version of the umbrella. Deliveries are slated to commence from December 2015 if all goes to plan