Abstract-In this work we propose a physical memristor/resistive switching device SPICE compact model, that is able to accurately fit both unipolar/bipolar devices settling to its current-voltage relationship. The proposed model is capable of reproducing essential device characteristics such as multilevel storage, temperature dependence, cycle/event handling and even the evolution of variability/parameter degradation with time. The developed compact model has been validated against two physical devices, fitting unipolar and bipolar switching. With no requirement of Verilog-A code, LTSpice and Spectre simulations reproduce distinctive phenomena such as the preforming state, voltage/cycle dependent random telegraph noise and device degradation.

Abstract-Since the memristor was first built in 2008 at HP Labs, no end of devices and models have been presented. Also, new applications appear frequently. However, the integration of the device at the circuit level is not straightforward, because available models are still immature and/or suppose high computational loads, making their simulation long and cumbersome. This study assists circuit/systems designers in the integration of memristors in their applications, while aiding model developers in the validation of their proposals. We introduce the use of a memristor application framework to support the work of both the model developer and the circuit designer. First, the framework includes a library with the best-known memristor models, being easily extensible with upcoming models. Systematic modifications have been applied to these models to provide better convergence and significant simulations speedups. Second, a quick device simulator allows the study of the response of the models under different scenarios, helping the designer with the stimuli and operation time selection. Third, fine tuning of the device including parameters variations and threshold determination is also supported. Finally, SPICE/Spectre subcircuit generation is provided to ease the integration of the devices in application circuits. The framework provides the designer with total control overconvergence, computational load, and the evolution of system variables, overcoming usual problems in the integration of memristive devices.

New CMOS technologies such as SOI or FinFET are expected to enhance SRAM radiation-induced soft error rates thanks to a reduction on the charge collected as the devices get smaller. In this work we analyze how the radiation hardening capabilities of SRAMs are affected when process variations are considered by simulating cells using a predictive FinFET technology.The results show that even if the average critical charge to which SRAM cells are vulnerable is enhanced by process variations, its widened spread leads to an increase of the soft error rate by more than 40% as the technology node is scaled down to 7nm.

No abstract

Abstract-Resistive switching memories (RRAM) are an attractive alternative to non-volatile storage and non-conventional computing systems, but their behavior strongly depends on the cell features, driver circuit and working conditions. In particular, the circuit temperature and the writing voltage scheme become critical issues, determining resistive switching memories performance. These dependencies usually force a design time trade-off among reliability, device endurance and power consumption, and therefore imposing non-flexible functioning schemes and limiting the system performance. In this paper we present a writing architecture that ensures the correct operation no matter the working temperature, and allows the dynamic load of application oriented writing profiles. Thus, taking advantage of more efficient configurations, the system can be dynamically adapted to overcome RRAM intrinsic challenges. Several profiles are analyzed regarding power consumption, temperature-variations protection and operation speed, showing speed-ups near to 700 x compared against other published drivers.