Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

RRAMSpec: A Design Space Exploration Framework for High Density Resistive RAM

 
: Mathew, Deepak M.; Chinazzo, André Lucas; Weis, Christian; Jung, Matthias; Giraud, Bastien; Vivet, Pascal; Levisse, Alexandre; Wehn, Norbert

:

Pnevmatikatos, D.N.:
Embedded Computer Systems: Architectures, Modeling, and Simulation: 19th International Conference, SAMOS 2019. Proceedings : July 7–11, 2019, Samos, Greece
Cham: Springer, 2019 (Theoretical Computer Science and General Issues 11733)
ISBN: 978-3-030-27562-4
ISBN: 978-3-030-27561-7
S.34-47
International Conference on Embedded Computer Systems - Architectures, Modeling, and Simulation (SAMOS) <19, 2019, Samos>
European Commission EC
732631; OPRECOMP
Open TransPREcision COMPuting
Englisch
Konferenzbeitrag
Fraunhofer IESE ()
RRAM; ReRAM; Crossbar; NVM

Abstract
Resistive RAM (RRAM) is a promising emerging Non-Volatile Memory candidate due to its scalability and CMOS compatibility, which enables the fabrication of high density RRAM crossbar arrays in Back-End-Of-Line CMOS processes. Fast and accurate architectural models of RRAM crossbar devices are required to perform system level design space explorations of new Storage Class Memory (SCM) architectures using RRAM e.g. Non-Volatile-DIMM-P (NVDIMM-P). The major challenge in architectural modeling is the trade-off between accuracy and computing intensity. In this paper we present RRAMSpec, an architecture design space exploration framework, which enables fast exploration of various architectural trade-offs in designing high density RRAM devices, at accuracy levels close to circuit level simulators. The framework estimates silicon area, timings, and energy for RRAM devices. It outperforms state-of-the-art RRAM modeling tools by conducting architectural explorations at very high accuracy levels within few seconds of execution time. Our evaluations show various trade-offs in designing RRAM crossbar arrays with respect to array sizes, write time and write energy. Finally we present the influence of technology scaling on different RRAM design trade-offs.

: http://publica.fraunhofer.de/dokumente/N-561656.html