MIT Researchers’ New Chip Design Swarm Significantly Improve Parallel Processing
For the past few years, cores have been added to processors in order to increase the performance output. As a general rule of thumb, more cores essentially mean more processors running at the same time.
There are a few applications which have been developed to utilize multiple cores simultaneously, but achieving that is not as simple as it sounds! Parallel program such as these require breaking up the program into multiple tasks and then forcing synchronization among these tasks, thus allowing shared data access.
To ease developers and improve performance, researchers at MIT have developed a chip design called Swarm. This new chip design Swarm essentially handles the synchronization all by itself which renders programming a whole lot easier.
The MIT researchers came up with six different Swarm versions of common algorithms that ran 3x to 18x faster. In one case, the algorithm ran 75x faster.
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.
As noted by OverclockersClub; Swarm essentially manages synchronization using its own time-stamping tasks based on priority which is something set by the programmer. This basically allows the chip to protect high priority tasks from getting incorrect data and decide whether a task can write to the shared memory or not.
What distinguishes Swarm from other multicore chips is that it has extra circuitry for handling that type of 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.
To read up more, head over to the MIT official site!