TCS-C Special Issue on New Computing Architectures - a Formal Glimpse
摘要截稿:
全文截稿: 2019-07-30
影响因子: 0.747
期刊难度:
CCF分类: B类
中科院JCR分区:
• 大类 : 计算机科学 - 4区
• 小类 : 计算机:理论方法 - 4区
Overview
The exponential performance improvements – Moore’s law, Dennard Scaling, etc. – of the general purpose ‘classic’ computing architecture era are coming to an end after thirty years. At the same time, computing practice is revealing new limits of current architectures that went unnoticed until recently: insecurity (hidden channels), DRAM weaknesses (rowhammer effect), energy inefficiency, etc. Dramatic changes are coming. Dropping out universality, like in the quantum annealing method used in D-Wave machines, and classical application-specific architectures advocated by 2017 Turing Award recipient John Hennessy, are symptomatic examples. New alternative computing architectures will appear: quantum, ‘post classic’, a renaissance of dataflow (EDGE), biological, etc.
The objective of this Theoretical Computer Science C Special Issue, titled “New Computing Architectures – a Formal Glimpse”, is to publish papers from diverse fields – ranging from computer architecture to mathematics, physics and theoretical computer science – to explore how mathematical, and, particularly, reworked formal foundational methods, could contribute to the design of reliable (re)constructions of new architectures with improved ‘algorithmic’ efficiency, performance and coverage. Mathematical, logical, formal tools and proofs are needed not only to design but also to give ``assurances” of the intended behaviour of new designs and machines. As John Rushby reminds us, ‘To make this notion of ``assurance” precise, we need to be careful about the terms we employ’, and, in using these mathematical tools, special care should be paid to avoid Gödelian, broken abstractions, or other known pitfalls.
Important questions will stand to be answered: Can these new architectures be carved so that we can mathematically reason about them? Can these tools be productively used to verify and ensure the reliability and viability of multiple conflicting security domains, multiple scenarios, and multiple user applications?