Dresden 2017 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

HL: Fachverband Halbleiterphysik

HL 52: Devices

HL 52.7: Talk

Wednesday, March 22, 2017, 11:15–11:30, POT 06

Cognitive computing with mem-circuits — •Alessandro Fumarola and Stuart Parkin — Max Planck Institute for Microstructure Physics, Halle (Saale), Sachsen-Anhalt, Germany

The time and energy spent transporting information between memory and processor in standard computers (across the so-called ’Von-Neumann bottleneck’) has become problematic for data-centric applications such as real-time image recognition and natural language processing[.

Non-VN systems, such as the human brain, are capable of full on-line learning and could be designed moving from reliable but binary devices to dense and analog (but less reliable) nanoscale elements. Previous work has been carried out, for example,with phase-change devices and other non-volatile memories, obtaining high machine learning performance in large-scale hardware/software systems.

In this work, we assess the suitability and advantages of other circuit elements with memory, namely memcapacitors, used as synaptic devices in such Non-VN configurations. Simulated power and performance benchmarks are provided. Possible solid-state implementation of binary and analog memcapacitors are also proposed and characterized.