Reconfigurable logic and multi-bit in-memory processing with ferroelectric memristors


IoT and in general embedded devices require energy-efficient solutions. Today frequently data transfer operations are the most energy-consuming operations in microcontrollers or microprocessors, in particular between memory or even caches on one side and processor cores on the other side at least if we compare them to simple arithmetic operations like e.g. an addition. Therefore, especially for embedded devices it makes sense to execute simple operations directly in-situ where the data or the operands are located or are generated, namely in the memory or directly at sensor nodes. Near- and in-memory-computing is in principle an answer to the high energy costs of data transfer operations.

An open and in the community frequently discussed question is how new devices like memristors can support this. In particular in-memory-processing, in which opposed to near memory-computing the storage element becomes an inherent part of the processing step, requires the basic memory element to be significantly modified.