System Bits: Sept. 27

Memory management scheme accommodates commercial chips
In an improvement to a memory management scheme presented last year in which MIT Computer Science and Artificial Intelligence Laboratory researchers unveiled what they said was a fundamentally new way of managing memory on computer chips — one that would use circuit space much more efficiently as chips continue to comprise more and more cores, or processing units — the team has devised an updated version to address existing chip designs.

In the previous version, the team said that in chips with hundreds of cores, the scheme could free up somewhere between 15 and 25 percent of on-chip memory, enabling much more efficient computation.

But their scheme assumed a certain type of computational behavior that most modern chips do not enforce. An updated version is more consistent with existing chip designs and has a few additional improvements.

The researchers reminded that the essential challenge posed by multicore chips is that they execute instructions in parallel, while in a traditional computer program, instructions are written in sequence. As such, computer scientists are constantly working on ways to make parallelization easier for computer programmers.

MIT researchers have found a new way of managing memory on computer chips that uses circuit space much more efficiently and is more consistent with existing chip designs.
(Source: Jose-Luis Olivares/MIT)

The initial version of the MIT researchers’ scheme, called Tardis, enforced a standard called sequential consistency. But with respect to reading and writing data — the only type of operations that a memory-management scheme like Tardis is concerned with — most modern chips don’t enforce even this relatively modest constraint. For example, a standard chip from Intel might assign the sequence of read/write instructions ABC to a core but let it execute in the order ACB, they said, and relaxing standards of consistency allows chips to run faster.

The researchers said Tardis uses chip space more efficiently than existing memory management schemes because it coordinates cores’ memory operations according to “logical time” rather than chronological time, and every data item in a shared memory bank has its own time stamp. Each core also has a counter that effectively time stamps the operations it performs. No two cores’ counters need agree, and any given core can keep churning away on data that has since been updated in main memory, provided that the other cores treat its computations as having happened earlier in time.

To enable Tardis to accommodate more relaxed consistency standards, the team said they simply gave each core two counters, one for read operations and one for write operations. If the core chooses to execute a read before the preceding write is complete, it simply gives it a lower time stamp, and the chip as a whole knows how to interpret the sequence of events.

Further, as different chip manufacturers have different consistency rules, much of a new paper on Tardis describes how to coordinate counters, both within a single core and among cores, to enforce those rules.

Wearable microscope measures through skin
In what could be an important tool for tracking various biochemical reactions for medical diagnostics and therapy, UCLA researchers working with a team at Verily Life Sciences have designed a mobile microscope that can detect and monitor fluorescent biomarkers inside the skin with a high level of sensitivity.

Researchers can detect spatial frequencies of a fluorescent image, which are then analyzed to sense the target fluorescence signal through the skin. (Source: Ozcan Research Group/UCLA)

The system weighs less than a one-tenth of a pound, making it small and light enough for a person to wear around their bicep, among other parts of their body. In the future, technology like this could be used for continuous patient monitoring at home or at point-of-care settings, the researchers said.

Given that fluorescent biomarkers are routinely used for cancer detection, and drug delivery, and release among other medical therapies, recent developments with biocompatible fluorescent dyes have emerged, creating new opportunities for noninvasive sensing and measuring of biomarkers through the skin. But detecting artificially added fluorescent objects under the skin is challenging since collagen, melanin and other biological structures emit natural light in a process called autofluorescence.

This microscope can monitor fluorescent biomarkers inside the skin. (Source: Ozcan Research Group/UCLA)

To measure the fluorescent dye, the wearable microscope used a laser to hit the skin at an angle. The fluorescent image at the surface of the skin was captured via the wearable microscope. The image was then uploaded to a computer where it was processed using a custom-designed algorithm, digitally separating the target fluorescent signal from the autofluorescence of the skin, at a very sensitive parts-per-billion level of detection.

The team believes this computational imaging framework might also be used in the future to continuously monitor various chronic diseases through the skin using an implantable or injectable fluorescent dye.

Quantum communications
According to Yale researchers, if you go to a particular spot outside the Oyster Bar in New York’s Grand Central Terminal, and murmur very softly into the corner, dozens of feet away, you can still be heard clearly.

This is known as a whispering gallery — a phenomenon in which sound waves of certain frequencies travel along curved surfaces. St. Paul’s Cathedral in London has the earliest known example. 

Now, researchers in the lab of Hong Tang, Yale’s Llewellyn West Jones Jr. Professor of Electrical Engineering & Physics, have taken the concept of the whispering gallery – using light waves instead of acoustic ones – and applied it to a device known as an optomagnonic resonator, which they believe could lead to a way to efficiently convert information from microwave photons to optical waves, which can be transmitted over optical fiber over very long distances, opening up new possibilities in quantum communications. 

Light interacts with magnons in a YIG sphere optomagnonic resonator and changes its polarization. (Source: Xufeng Zhang, Yale)

Magnons are the smallest unit of measurement for a magnetic spin excitation, and the researchers asserted they hold great promise as an information carrier in quantum systems partly because they can simultaneously interact with multiple different types of information carriers including microwave photon, optical photon, acoustic phonon, etc. They said this is important since one of the greatest challenges in developing complex hybrid systems is realizing an efficient microwave-to-optical conversion to get the benefits of both. 

The team said the next step for the innovation is to improve the magnon-to-photon conversion efficiency by means such as reducing the YIG sphere size to further concentrate the optical and magnon modes, and doping the material to increase the interaction coefficient.