Researchers are experimenting with different forms of memory that might be better suited for AI, with indium gallium zinc oxide (IGZO) being the latest to gain attention.
IGZO-based 2-transistor 1-capacitor (2T1C) technology is typically found in display screens, but research organization imec has identified its potential for analog in-memory computing (AIMC).
The AIMC approach addresses the limitations of traditional digital computing, specifically speed and energy efficiency, by executing computing tasks within memory itself using analog technology. This minimizes power consumption and accelerates computational speed.
Denser memory array
The key advantage lies in parallel processing and storing data in analog format within memory, offering a faster, more efficient and energy-saving mode of calculation. In essence, the memory itself becomes part of the calculation process, eliminating the need for data transfers between separate drives.
IGZO DRAM cells show great promise for analog in-memory computing due to their significantly reduced standby power consumption. Additionally, IGZO transistors can be processed at the back end of line (BEOL) of the chip, allowing them to be placed on top of peripheral circuitry located at the front end of line (FEOL). This results in a denser memory array with no FEOL footprint.
At the recent International Memory Workshop (IMW) 2023, imec teams addressed some of the remaining challenges, strategies for optimizing gain cell retention time, and demonstrations of successful MAC operation in an array configuration.
It wasn't the only company to talk about the technology either, as Samsung also shared its research there.
You can read more about the topic on the imec site, but the researchers conclude that IGZO-based 2T1C and 2T0C (a capacitor-less variant) gain cells show exceptional properties for AIMC. Compared to traditional SRAM-based technology, they offer superior energy efficiency and computational density for machine learning applications, particularly during the inference phase. 2T0C cells stand out even more in area efficiency.