Models of mem-capacitors - computing devices with memory based on graphene nanocomposites are developed in the laboratory of the university (Russia). This class of devices has an architecture with significantly greater speed and lower power consumption compared to traditional ones. It allows avoiding the memory-processor bandwidth problem of the device. The laboratory is looking for partners for technical cooperation, licensing or commercial agreements with technical assistance.
Princeton architecture (or von Neumann architecture) is ubiquitous in modern computing devices. One of the most serious bottlenecks of this architecture is the limitation of bandwidth between the processor and memory compared to the amount of memory. Due to the fact that the processor speed and memory size increase much faster than the bandwidth between them, the bottleneck has become a big problem, the seriousness of which increases with each new generation of processors.
To solve this problem, university staff (Novosibirsk) obtained a new class of electronic devices with memory, the so-called capacitors with memory or memory capacitors, with the aim of introduction. Using various types of graphene nanocomposites, the possibilities of creating such devices were investigated, as well as a fundamental understanding of the possibilities of their functioning was achieved. As a result of research, models of devices were developed at the microscopic level and in conditions of a working chain of elements.
The implementation of these devices allows one to create a foundation for developing the elemental base – mem-capacitors based on graphene membranes and ferroelectrics - computing devices of the new generation, which have significantly higher performance and lower energy costs compared to traditional ones. Algorithms of logical operations implemented using memory elements let one push the boundaries of the computer technologies application field and overcome the problems of currently existing computer architectures.
The introduction of these devices contributes to the creation of artificial neural networks - mathematical models, as well as their software and hardware implementations, built on the principle of organization and functioning of biological neural networks. The implementation of neural networks is used in a wide range of areas such as economics and business, medicine and health care, communications, the Internet, production automation, political science, and sociology technology, geological exploration, and others.
The laboratory is looking for partners for R & D, licensees for the provision of a non-exclusive license, industrial partners for the introduction of the developed technology into production.
Innovative aspects consist of the process of creation devices with
- significantly higher speed,
- lower energy costs
compared to traditional ones.
The use of mem-capacitors expands the field of applicability of computer technology and allows to overcome the problems of currently existing computer architectures (for example, von Neumann architecture).
On the basis of these devices, work on the construction and operation of artificial neural networks, which are now widely used in the field of machine learning and data analysis, is accelerated.
Current development status
Device models received.