A team of top international scientists is working to create memory chips that double as processing elements, a breakthrough that could pave the way for novel computing approaches.
Researchers from Nanyang Technological University, Singapore (NTU Singapore) worked in collaboration with Germany’s RWTH Aachen University and Forschungszentrum Jülich to develop a prototype circuit using a memory technology called Redox-based resistive switching random access memory (ReRAM). This state-of-the-art non-volatile memory promises long-term storage capacity and low energy usage, which has led semiconductor companies like SanDisk and Panasonic to invest heavily in the technology.
In their paper, “Multistate Memristive Tantalum Oxide Devices for Ternary Arithmetic,” the authors note:
“Redox-based resistive switching random access memories (ReRAMs) are considered as one of the most promising emerging non-volatile memory technologies. The devices can be scaled down to 5 nm, offer endurance up to 1012 cycles, 10 years retention and fast read/write speed of below 200 ps. […] Up to 8 multi-states have been shown, allowing the storage of up to three binary digits in a single cell.”
The research sets the stage for memory chips that can also perform calculations.
“The current paper reports the first implementation of modular arithmetic using multi-state ReRAM devices, which is fully crossbar array compatible in conjunction with a selector device,” the authors write. “So far, most previous memristive circuit studies are based on over simplistic memristor models, we hereby use real memristive devices fabricated in word structures to verify the proposed functionality.”
The team developed an algorithm that calculates the carries and sums directly in the ReRAM devices, which store the results until they are read out. “Initially, all the devices in a wordline are initialized, i.e. written to the LRS [low resistance state]. Starting from this state the sum bit of significance 0 (s0) can be directly calculated in the device of significance 0 while the other devices are calculating the first output carry c1. The actual sum or carry calculating devices are shifted for each significance one device to the left.”
Where traditional binary computers are relegated to storing and processing data as 0 or 1, ReRAM devices use a ternary number system, comprised of 0, 1 and 2.
In the binary system used by today’s computers, all information must be translated into a string of zeros and ones.
Assistant Professor Anupam Chattopadhyay with NTU’s School of Computer Science and Engineering explains, one of the principal researchers on the project, said, “This is like having a long conversation with someone through a tiny translator, which is a time-consuming and effort-intensive process. We are now able to increase the capacity of the translator, so it can process data more efficiently.”
“By using a ternary number system, the amount of devices and cycles can be reduced significantly,” observe the paper’s authors. “In contrast to two-state devices, multistate devices provide better radix economy with the option for further scaling. Therefore, establishing multi-state ReRAM for non-volatile memory opens the door to novel storage and in-memory computer arithmetic options.”
As for real world applications that could benefit from the novel approach, the authors write that “low-variance multi-level ReRAM could play a key role for implementation of public-key cryptography and error-correcting codes in smart devices.”
The Scientific Reports paper “Multistate Memristive Tantalum Oxide Devices for Ternary Arithmetic” was authored by Wonjoo Kim and Vikas Rana of Forschungszentrum Jülich; NTU’s Anupam Chattopadhyay; and Anne Siemon, Eike Linn, and Rainer Waser of RWTH Aachen University. The online version was published on November 11, 2016.