American IT engineers from the University of California at Riverside have found a way to significantly speed up the operation of computer systems and at the same time reduce energy consumption without the need to improve equipment. The study was published on: Web site Institute of Electrical and Electronics Engineers (IEEE).
The method is based on a process called concurrent and heterogeneous multithreading (SHMT). It uses various types of processors found in modern computers – graphic, central and tensor (for artificial intelligence technologies).
The test setup included an ARM Cortex-A57 CPU, an Nvidia GPU, and a Google Edge Tensor Processor. With concurrent and heterogeneous multithreading, sample code execution was 1.95 times faster and energy consumption was reduced by 51%.
“Established programming models focus on using only the most efficient processors for each code region and do not adequately leverage the processing power of heterogeneous computers,” the researchers wrote in their paper.
Scientists have acknowledged that they need to overcome significant challenges related to separating the computing tasks to be performed by different types of processors and then putting everything back together without any slowdown. Therefore, SHMT technology will not be widely implemented in the near future.
Previous scientists stated It’s about building a chip for light-speed AI computing.
What are you thinking?
Source: Gazeta
Jackson Ruhl is a tech and sci-fi expert, who writes for “Social Bites”. He brings his readers the latest news and developments from the world of technology and science fiction.