Answers are grounded in this paper and cite their sources.
This paper presents an analog dot-product compute-in-memory (CIM) using quantum dot-nonvolatile random access memory (QD-NVRAM). In this approach, the gate of the QD-NVRAM is modulated to control the spatial location of the charges in the floating quantum dot layer, thus generating unique I–V characteristics. Here, the gate voltage is modulated using a complementary p-quantum dot channel (QDC)-field effect transistor (FET) and an n-QDC-FET either by charging or discharging the gate of QD-NVRAM. In this QD-NVRAM, the additional drain region has direct access to the floating gate layer of the quantum dot access channel, thus increasing the speed of the erase and/or write operation. The QD-NVRAM-based computing-in-memory could improve the latency and energy efficiency of computation-intensive devices in artificial intelligence applications.
No takes yet. Share an insight, caveat, or question.
Gudlavalleti et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: