Sie befinden Sich nicht im Netzwerk der Universität Paderborn. Der Zugriff auf elektronische Ressourcen ist gegebenenfalls nur via VPN oder Shibboleth (DFN-AAI) möglich. mehr Informationen...
Implementation of analog RRAM technology for neuromorphic computing has faced unique challenges in terms of endurance and variability as compared to the conventional digital memories. In this work, we demonstrate the dual oxide layer RRAM devices with analog performance of low working current, and study the trade-off switching speed, and the on/off ratio. We evaluate the endurance characteristics and define the correlation of endurance degradation with variability behavior. Compromised specifications of switching speed, switching voltage, and working resistance range are suggested to achieve integrative performance for the hardware implementation of neuromorphic network.