System Bits: June 21

Faster running parallel programs, one-tenth the code
MIT researchers reminded that computer chips have stopped getting faster and that for the past 10 years, performance improvements have come from the addition of cores. In theory, they said, a program on a 64-core machine would be 64 times as fast as it would be on a single-core machine but it rarely works out that way. Most computer programs are sequential, and splitting them up so that chunks of them can run in parallel causes all kinds of complications.

However, researchers from MIT’s Computer Science and Artificial Intelligence Laboratory have devised a new chip design they call Swarm, which they said should make parallel programs not only much more efficient but easier to write.

“Multicore systems are really hard to program,” says Daniel Sanchez, an assistant professor in MIT’s Department of Electrical Engineering and Computer Science. “You have to explicitly divide the work that you’re doing into tasks, and then you need to enforce some synchronization between tasks accessing shared data. What this architecture does, essentially, is to remove all sorts of explicit synchronization, to make parallel programming much easier.”
(Source: MIT)

In simulations, the researchers compared Swarm versions of six common algorithms with the best existing parallel versions, which had been individually engineered by seasoned software developers. The Swarm versions were between three and 18 times as fast, but they generally required only one-tenth as much code — or even less. And in one case, Swarm achieved a 75-fold speedup on a program that computer scientists had so far failed to parallelize.

Multicore systems are really hard to program, and the work has to be explicitly divided into tasks, and then some synchronization between tasks accessing shared data needs to be enforced. This architecture essentially removes all sorts of explicit synchronization, to make parallel programming much easier.

The researchers said what distinguishes Swarm from other multicore chips is that it has extra circuitry for handling prioritization. It time-stamps tasks according to their priorities and begins working on the highest-priority tasks in parallel. Higher-priority tasks may engender their own lower-priority tasks, but Swarm slots those into its queue of tasks automatically.

Manipulating light, matter
Following theoretical simulations by scientists from Imperial College London and Kings College London, University of Cambridge researchers have mixed a molecule with light between gold particles, creating a new way to manipulate the physical and chemical properties of matter. This mixing is called strong coupling.

The team noted that light and matter are usually separate and have distinct properties, but molecules of matter can emit particles of light called photons. Normally, emitted photons leave the molecule and the two do not mix again.

Now, scientists have trapped a single molecule in such a tiny space that when it emits a photon, the photon cannot escape. This produces an oscillation of energy between the molecule and the photon, creating a mixing of the properties of matter and light.

Illustration of a molecule coupled with light in a gold
nanopore.
(Source: Imperial College London)

This unusual interaction of a molecule with light will provide new ways to manipulate the physical and chemical properties of matter, and could be used to process quantum information, aid in the understanding of complex processes at work in photosynthesis, or even manipulate the chemical bonds between atoms.

These findings could be useful in quantum technologies since light carries quantum information, so strong coupling could be used to copy the information over to matter and back.

Smart threads for sharing data
Researchers at UC Berkeley have created technology that coats individual threads in fabric with thermochromic paint that heats up and gradually changes the thread colors when given a jolt of electricity so that users could convey mood or other information on their clothing.

The team worked with Google’s Advanced Technology and Projects group, an in-house technology incubator in connection with ATAP’s Project Jacquard, a platform for embedding sensors and feedback devices in fabrics and clothing in ways that seem natural and comfortable.

The platform encompasses techniques for creating fashion fabrics with conductive fibers woven into them, plus small, flexible computing components and feedback devices (such as haptics or LEDs), along with software that application programming interfaces can use to exchange data with the garment.

The research team developed seven crocheted and woven fabric swatches that have been tested with fashion designers and others.

Woven seven-segment grid displays different numbers
(Source: UC Berkeley)