Integrated Memristor-Based Computer Architectures

基于忆阻器的集成计算机架构

• 批准号: 389549790
• 负责人: Professor Dr.-Ing. Dietmar Fey
• 金额: --
• 依托单位: Lehrstuhl für Technische Elektronik
• 依托单位国家: 德国
• 项目类别: Research Grants
• 财政年份: 2017
• 资助国家: 德国
• 起止时间: 2016-12-31 至 2020-12-31
• 项目状态: 已结题
• 来源: https://gepris.dfg.de/gepris/projekt/389549790?language=en
• 关键词: Integrated; Memristor; Based; Computer; Architectures

In recent years, in the development of memristive devices significant progress has been attained, so that the commercialization of new background memory systems based on this technology has already begun. In present research memristive devices are only used for storing data but not for the processing of data inside a CPU. However, from an energy point of view and based on the memristors non-volatile behavior, memristive devices should also be of interest for data processing, especially for CPUs in IoT-devices utilizing a volatile power supply. Therefore, the aim of this project is to use memristor technology in modern computer architectures; especially for the integration of memristive devices in a CPU for embedded applications. For a further reduction of the energy consumption also the ability of storing more than one bit in a single memristor cell can be used to enable the realization of tenary logic, which requires multi-bit-storage in a memristive register. Tenary logic will result in a redundant number representation, which allows to execute an adding during a constant time independent of the word length. This also results in a decreased power consumption.Due to the compatibility of the memristors technology with standard CMOS processing, an integration of memristors will be possible by postprocessing the chip. However, currently an automated and tool-aided design flow for memristive devices, especially for digital designs, is not available. Instead, dedicated mixed signal circuits have to be developed to enable the use of memristors also in the digital world. Regarding the realization of the tenary logic the problem has to be solved to what extent ceaseless conversions between the analog and the digital domain make sense. Therefore, in addition to the integration of memristors a further goal of this project is the development of a tenary processor architecture utilizing memristive devices. For the realization of such an architecture, e.g., basic mixed-signal blocks can be used. Consequently, memristors will not only be used as pure storage devices but will be part of the data processing within a CPU.The computer architectures will be realized as a prototype IC and verified. The integration of the memristive devices is done through a special BEOL processing in which the memristors will be placed at the top chip layer. Using appropriate benchmarks a qualitative and quantitative evaluation of the developed architectures will be performed to assess the usefulness of memristors for data processing in modern CPUs. A final evaluation shall result in the identification of possible applications fields of the designed architecture.

近年来，在忆阻器件的开发中已经取得了重大进展，使得基于该技术的新的背景存储器系统的商业化已经开始。在目前的研究中，忆阻器件仅用于存储数据，而不用于CPU内部的数据处理。然而，从能量的角度来看，并且基于忆阻器的非易失性行为，忆阻器件也应该对数据处理感兴趣，特别是对于利用易失性电源的IoT设备中的CPU。因此，本项目的目的是在现代计算机架构中使用忆阻器技术;特别是在嵌入式应用的CPU中集成忆阻器件。为了进一步减少能量消耗，还可以使用在单个忆阻器单元中存储多于一个比特的能力来实现十进制逻辑，这需要忆阻寄存器中的多比特存储。十进制逻辑将导致冗余数表示，这允许在恒定时间内执行加法，而与字长无关。由于忆阻器技术与标准CMOS工艺的兼容性，通过对芯片进行后处理可以实现忆阻器的集成。然而，目前还没有用于忆阻器件的自动化和工具辅助的设计流程，特别是用于数字设计的自动化和工具辅助的设计流程。相反，必须开发专用的混合信号电路，以使忆阻器也能在数字世界中使用。关于十元逻辑的实现，必须解决的问题是，模拟域和数字域之间的不断转换在多大程度上是有意义的。因此，除了忆阻器的集成，该项目的另一个目标是利用忆阻器件开发十元处理器架构。为了实现这样的架构，例如，可以使用基本的混合信号块。因此，忆阻器将不仅被用作纯存储设备，而且将成为CPU内数据处理的一部分。计算机架构将被实现为原型IC并得到验证。忆阻器件的集成是通过特殊的BEOL工艺完成的，其中忆阻器将被放置在顶部芯片层。使用适当的基准的定性和定量评估的开发架构将进行评估的有用性忆阻器的数据处理在现代CPU。最终评估应确定所设计架构的可能应用领域。