Processors in devices use a layered set of memory caches to provide for a fast processing by fetching for commonly used data. It has been used for a very long time now and a new technology is on the verge of replacing it. Researchers at CSAIL, MIT have developed a cache system that creates cache locations on the spot for specific apps. The system is being called the Jenga and the system knows the physical location of each memory bank.
Jenga can calculate how to store data to reduce travel time which reduces the lag considerably. The system works efficiently whether an app would benefit from multiple levels of cache or from a single big cache.
A simulation also revealed that Jenga can speed up a 36-core chip up to 30 percent and could also use 85 percent less power. This means that processors with multi-cores wouldn't require a lot of power from a device. This feature of the system makes it an ideal add-on for mobile devices such as laptops and smartphones.
Jenga has just seen a simulation test and there is time before the system is developed for real-time testing. It would take even more time for manufacturers to adopt the cache system and incorporate it into their products. Another issue at hand is Jenga's performance across several processors with varying core count. Jenga is a promising piece of system enhancement and with time it will only get better with time.
Large scale manufacturers such as Qualcomm may adopt the idea sooner and work on it to make its Socs power efficient and faster.