Designing Reliable Systems Using Unreliable Components

项目编号：0737655标题：使用不可靠的材料设计可靠的系统PI姓名：Chris Myers机构：犹他大学保持摩尔定律和每年生产更快的计算机芯片变得越来越困难。 这在很大程度上是由于晶体管越小，它们就越不可靠。 随着我们进入纳米技术时代，这个问题将变得更加严重。 目前处理这些问题的方法非常耗电。 然而，有证据表明，有更好的方法，即生物系统在具有极高噪声和不可靠组件的情况下实现惊人的功率效率。例如，人类内耳能够每秒执行相当于10亿次浮点运算，而仅消耗14微瓦的功率。 这一性能与处理器消耗约50瓦功率的游戏机相当。 因此，该项目的目标是从生物系统中汲取灵感，以开发新的设计方法和工具，用于使用不可靠的组件设计可靠的系统。 这项工作的目标是设计由传统硅晶体管和新的纳米技术器件构成的系统。 此外，我们计划将这些技术应用于合成生物设备。 这一研究领域有望设计细菌来构建用于医疗用途的复杂天然产物，有毒废物管理领域，甚至检测和杀死肿瘤细胞。

Project ID: 0737655Title: SGER Designing Reliable Systems Using Unreliable ComponentsPI name: Chris MyersInstitution: University of UtahIt is becoming increasingly difficult to maintain Moore's Law and produce faster computer chips each year. This is largely due to the fact that as transistors get smaller, they become more unreliable. As we move into the nanotechnology era, this problem will become even worse. Current approaches to deal with these problems are extremely power hungry. There is, however, proof that there are better approaches in that biological systems achieve amazing power efficiency with extremely noisy and unreliable components.For example, the human inner ear is capable of performing the equivalent of one billion floating point operations per second while consuming only 14 microwatts of power. This performance is comparable to that of a game console with a processor that burns about 50 watts of power. Therefore, the goal of this project is to draw inspiration from biological systems in order to develop new design methods and tools for designing reliable systems using unreliable components. This work will target the design of systems constructed from both traditional silicon transistors as well as new nanotechnology devices. In addition, we plan to apply these techniques to synthetic biological devices. This area of research holds the promise of designing bacteria to construct complex natural products for medical use, areas of toxic waste management, or even detect and kill tumor cells.