The concept was described in a paper that appeared recently in the journal Physical Review X. The research was performed by a Purdue University team led by Supriyo Datta, the Thomas Duncan Distinguished Professor of Electrical and Computer Engineering.

Digital computers use strings of zeroes and ones called bits to represent information in a binary code. Quantum computers use q-bits tapping into a phenomenon known as superposition that allows a bit to be both zero and one at the same time. PSL uses a p-bit which is either zero or one at a given time, but it fluctuates rapidly between the two with probabilities that can be controlled.

“The fact that a p-bit is stochastic sounds like a bad thing, but we have shown earlier that p-transistors based on p-bits can be very useful for problems involving optimization and Bayesian inference. In this paper we showed that they can even be useful for precise arithmetic functions like adders and multipliers,” said postdoctoral research associate Kerem Camsari, the first author of the paper. “Moreover, it is invertible: an adder circuit can operate as a subtractor, and a multiplier circuit can operate as a divider or a factorizer — and this can be important since many computational problems are known to be harder when going in one direction than the other.”

The research was funded by the Center for Spintronic Materials, Interfaces and Novel Architecture, or C-SPIN, as well as the Purdue-led Institute for Nanoelectronics Discovery and Exploration (INDEX), where researchers are working on experiments implementing p-bits and p-transistors.

The paper, authored by Camsari; graduate students Rafatul Faria and Brian M. Sutton; and Datta, is available online at https://journals.aps.org/prx/abstract/10.1103/PhysRevX.7.031014. More information is available by contacting the paper’s lead author, Camsari, at kcamsari@purdue.edu.

Source: Purdue Newsroom