A new experimental approach developed by Hung-Wei Tseng, an associate professor of electrical and computer engineering at UC Riverside, addresses this problem by removing the sequential processing of information. Called Simultaneous and Heterogenous Multithreading (SHMT), this approach, as the name suggests, allows simultaneous information processing.

Along with Kuan-Chieh Hsu, a graduate student in computer science at UC Riverside, Tseng demonstrated his SHMT framework using a multi-core ARM processor, an NVIDIA GPU, and a Tensor Processing Unit hardware accelerator. When tested, the approach delivered a 1.96 times increase in processing speed and a 51 percent reduction in power consumption.