Descripción
|
|
---|---|
Resistive switching memories (RRAM) are one of the most promising alternatives for non-volatile storage and non-conventional computing systems. However, their behavior, and therefore their reliability is limited by technology intrinsic constraints. Standard CMOS reliability analyses do not take into account RRAM related misbehaviors. Consequently, new and more thorough characterization approaches are needed. Even more important, as RRAM is proposed to become a key piece in aerospace solutions, new radiation and temperature analyses should be also considered in reliability oriented methodologies. This work presents a solution that completely characterizes RRAM and CMOS hybrid circuits under the combined effects of both technology and environmental error sources. The analyses strategy is based on three pillars: the definition of suitable models, the application of user-defined metrics to measure both circuit reliability and performance, and the efficient definition of the design space. These concepts are used by a powerful simula- tion framework, achieving automatic characterization of RRAM based circuits by simultaneously considering multiple error sources. As a case of study, a thorough analysis of an RRAM read driver, including RRAM lifetime, circuit temperature, CMOS and RRAM variability and radiation ?both accumulated dose and single particle impacts? highlights the proposed approach capabilities. | |
Internacional
|
Si |
JCR del ISI
|
Si |
Título de la revista
|
Ieee Transactions on Nanotechnology |
ISSN
|
1536-125X |
Factor de impacto JCR
|
1,702 |
Información de impacto
|
Datos JCR del año 2015 |
Volumen
|
16 |
DOI
|
10.1109/TNANO.2017.2697311 |
Número de revista
|
Volume 16, |
Desde la página
|
1 |
Hasta la página
|
9 |
Mes
|
JULIO |
Ranking
|