“Words lie in our way”

The central claim of computationalism is generally taken to be that the brain is a computer, and that any computer implementing the appropriate program would ipso facto have a mind. In this paper I argue for the following propositions: (1) The central claim of computationalism is not about computers, a concept too imprecise for a scientific claim of this sort, but is about physical calculi (instantiated discrete formal systems). (2) In matters of formality, interpretability, and so forth, analog computation and digital computation are not essentially different, and so arguments such as Searle's hold or not as well for one as for the other. (3) Whether or not a biological system (such as the brain) is computational is a scientific matter of fact. (4) A substantive scientific question for cognitive science is whether cognition is better modeled by discrete representations or by continuous representations. (5) Cognitive science and AI need a theoretical construct that is the continuous analog of a calculus. The discussion of these propositions will illuminate several terminology traps, in which it's all too easy to become ensnared.     

自然计算研究进展

随着量子计算、DNA计算、免疫计算、进化计算和神经计算等新型计算机制的兴起,传统的计算正扩展为自然计算,因为这些新计算机制的产生都由自然系统启发而来.为了发展和应用自然计算技术,总结了自然计算的发展简史.然后,提出了计算机、自动控制等系统的两种体系结构,分析了自然计算的新特征,包括自然法则、交叉性、革命性和复杂性,并给出了自然计算的复杂性分析定理.改进了自然计算的映射模型,提出了自然计算的算法流程.最后,探讨了自然计算的研究意义和应用.总之,自然计算是活跃的重要研究方向,定会推动许多学科的发展.     

Hybrid Computer Systems

Today's increasing emphasis on simulation of continuous dynamic systems lends a special significance to this issue of Computer. Hybrid computers, with their tremendous computing speed resulting from the all-parallel configuration of their analog subsystem, are particularly suited for dynamic system simulation. Indeed, back in the 1950's analog computers were the only game in town when it came to real-time simulation of flight vehicles. Digital computers of that era were simply too slow in the numerical solution of differential equations.     

采用PETSc的有限元并行计算实现与优化

可移植可扩展科学计算工具箱PETSc提供了高性能求解偏微分方程组的大量对象和解法库,基于此进行结构有限元并行计算,可降低难度和成本。给出了基于PETS的结构有限元并行计算实现方法,包括有限元方程组的并行形成和并行求解的实现。根据PETSc的特点,提出了提高计算性能的优化措施,即数据局部化和存储预分配。数值实验表明实现方法可行,优化措施效果明显。     

一种异步BSP模型及其程序优化技术

基于BSP模型，该文提出了异步计算模型（CSA－BSP）。该模型更准确地描述了并行机的性能参数，引导用户编写高效率的并行程序，在CSA－BSP模型下，两个进程异步执行的位置至多相差p-1个超步；基于程序的执行时间，作者分析了BSP、A－BSP和CSA－BSP程序的效率，得出CSA－BSP程序的效率是最高的，在曙光并行机上，用“红黑格法”和“矩阵乘法”进行了验证，和BSP模型相比，这两个CSA－BSP程序的效率分别提高20％和37％；同时，其进程执行时间和最大可以降低8％，因此，按照CSA－BSP模型编程对于提高程序效率和改善系统的吞吐率，都有良好的效果。     

高分辨率数值计算研究

高分辨率计算是高置信度计算中一个极其重要而复杂的研究问题。相对传统的数值计算,高分辨率计算对计算机系统和应用程序(物理建模、参数、计算方法和算法等)提出了很高的要求。并行计算机的发展为大规模科学计算,特别是数值计算分辨率的提高提供了条件。同时,数值计算分辨率的提高也对计算机的计算能力、计算方法、物理建模和参数等提出了新的、更高的要求。本文以一个二维流体力学程序计算平面爆轰问题为例,研究在计算分辨率提高时初始起爆区域、时间步长、网格构造、人为粘性、计算机模拟误差、计算量增长等方面出现的问题,提出了相应的解决办法,提高了计算的精确度。     

Parallelism in multigrid methods: How much is too much?

Multigrid methods are powerful techniques to accelerate the solution of computationally-intensive problems arising in a broad range of applications. Used in conjunction with iterative processes for solving partial differential equations, multigrid methods speed up iterative methods by moving the computation from the original mesh covering the problem domain through a series of coarser meshes. But this hierarchical structure leaves domain-parallel versions of the standard multigrid algorithms with a deficiency of parallelism on coarser grids. To compensate, several parallel multigrid strategies with more parallelism, but also more work, have been designed. We examine these parallel strategies and compare them to simpler standard algorithms to try to determine which techniques are more efficient and practical. We consider three parallel multigrid strategies: (1) domain-parallel versions of the standard V-cycle and F-cycle algorithms; (2) a multiple coarse grid algorithm, proposed by Fredrickson and McBryan, which generates several coarse grids for each fine grid; and (3) two Rosendale algorithm, which allow computation on all grids simultaneously. We study an elliptic model problem on simple domains, discretized with finite difference techniques on block-structured meshes in two or three dimensions with up to 10⁶ or 10⁹ points, respectively. We analyze performance using three models of parallel computation: the PRAM and two bridging models. The bridging models reflect the salient characteristics of two kinds of parallel computers: SIMD fine-grain computers, which contain a large number of small (bitserial) processors, and SPMD medium-grain computers, which have a more modest number of powerful (single chip) processors. Our analysis suggests that the standard algorithms are substantially more efficient than algorithms utilizing either parallel strategy. Both parallel strategies need too much extra work to compensate for their extra parallelism. They require a highly impractical number of processors to be competitive with simpler, standard algorithms. The analysis also suggests that the F-cycle, with the appropriate optimization techniques, is more efficient than the V-cycle under a broad range of problem, implementation, and machine characteristics, despite the fact that it exhibits even less parallelism than the V-cycle. Research at Princeton University partially supported by the National Science Foundation, Grant No. CCR-8920505, and the Office of Naval Research, Contract No. N0014-91-J-1463.     

Measuring perceptual and motivational facets of computer control: the development and validation of the computing control scale

The Computing Control Scale (CCS), a new factor analytically derived psychometric instrument is developed. The CCS consists of computing autonomy and computing need for control (NControl) subscales. Computing autonomy represents a composite of confidence in controlling computers and self-reliance when using computers. Computing NControl is considered to represent a domain-specific analog of Burger's (1992) global desire for control construct. The factor structure of the instrument is shown to be replicable. Also, the two subscales are shown to be reliable and to exhibit construct validity in terms of their differential relationships with other concepts such as computer comfort – anxiety, computer addiction and non-domain-specific desire for control. In addition, the data collected shows that few people attribute computing-related outcomes to luck or chance and indicates that the vast majority of people believe that in principle such outcomes are within their control. It is therefore concluded that attempts to measure computing-specific locus of control using a factor analytically derived instrument may not be viable.     

连续系统的数字仿真方法

近年来,已逐渐用数字计算机代替模拟计算机或混合计算机来实现连续系统的仿真.连续系统的数字仿真的显著特点是:便于人机对话,求解精度高和重复性好,容易编排程序和输出计算结果.但是,为了避开编程序的细节和把精力集中在系统仿真上,系统仿真者希望有一个方便和灵活的仿真语言.本文介绍一种面向方程的BASIC仿真语言,它允许用户直接用一阶微分方程组的数学表达式写入,并能在任何一台配置BASIC语言的数字机上运算.使用时只要遵循规定的简单格式而不用去熟悉BASIC语言的细节.文中附有阐述用法的计算实例.     

基于NUMECA FINE/Turbo的并行计算测试

为具体了解CFD软件NUMECA FINE/Turbo的并行计算性能,良好把握后续的科研工作进度,分别研究在激活超线程情况下单节点计算与多节点并行计算以及CPU在激活超线程前、后计算速度的差异.结果表明:在多节点并行计算时,计算速度与实际参加并行计算的CPU物理核心数量成正比;在激活超线程的情况下,并行计算节点数在超过实际物理核心数后明显降低计算速度的提升.     

A Quantitative/Method of Speed Comparison Between Analog/Hybrid and Digital Computers

For many years it has been claimed that analog and hybrid computers have a sizeable speed advantage over digital computers in the solution of differential equations. This speed advantage has, in fact, been documented for a number of specific studies, where direct comparisons with all-digital solutions have shown reduced run times and resulting cost savings using analog or hybrid computers. The purpose of this paper is to present a quantitative method of speed comparison that is quite general and allows direct speed comparisons between specific analog and digital computers. The method is based on previous reports written by the authors.^1,2Ability to make such speed comparisons is particularly important at this time when faster and faster digital computers are being marketed at generally lower prices.     

computation,among other things,is beneath us

What's computation? The received answer is that computation is a computer at work, and a computer at work is that which can be modelled as a Turing machine at work. Unfortunately, as John Searle has recently argued, and as others have agreed, the received answer appears to imply that AI and Cog Sci are a royal waste of time. The argument here is alarmingly simple: AI and Cog Sci (of the “Strong” sort, anyway) are committed to the view that cognition is computation (or brains are computers); butall processes are computations (orall physical things are computers); so AI and Cog Sci are positively silly. I refute this argument herein, in part by defining the locutions ‘x is a computer’ and ‘c is a computation’ in a way that blocks Searle's argument but exploits the hard-to-deny link between What's Computation? and the theory of computation. However, I also provide, at the end of this essay, an argument which, it seems to me, implies not that AI and Cog Sci are silly, but that they're based on a form of computation that is well “beneath” human persons.     

What is computation?

Three conditions are usually given that must be satisfied by a process in order for it to be called a computation, namely, there must exist a finite length algorithm for the process, the algorithm must terminate in finite time for valid inputs and return a valid output and, finally, the algorithm must never return an output for invalid inputs. These three conditions are advanced as being necessary and sufficient for the process to be computable by a universal model of computation. In fact, these conditions are neither necessary nor sufficient. On the one hand, recently defined paradigms show how certain processes that do not satisfy one or more of the aforementioned properties can indeed be carried out in principle on new, more powerful, types of computers, and hence can be considered as computations. Thus, the conditions are not necessary. On the other hand, contemporary work in unconventional computation has demonstrated the existence of processes that satisfy the three stated conditions, yet contradict the Church–Turing thesis and, more generally, the principle of universality in computer science. Thus, the conditions are not sufficient.     

Local and global models of physics and computation

Classical computation is essentially local in time, yet some formulations of physics are global in time. Here, I examine these differences and suggest that certain forms of unconventional computation are needed to model physical processes and complex systems. These include certain forms of analogue computing, massively parallel field computing and self-modifying computations.     

Physical Computation: How General are Gandy’s Principles for Mechanisms?

What are the limits of physical computation? In his ‘Church’s Thesis and Principles for Mechanisms’, Turing’s student Robin Gandy proved that any machine satisfying four idealised physical ‘principles’ is equivalent to some Turing machine. Gandy’s four principles in effect define a class of computing machines (‘Gandy machines’). Our question is: What is the relationship of this class to the class of all (ideal) physical computing machines? Gandy himself suggests that the relationship is identity. We do not share this view. We will point to interesting examples of (ideal) physical machines that fall outside the class of Gandy machines and compute functions that are not Turing-machine computable.     

HLognGP: A parallel computation model for GPU clusters

Parallel computation model is an abstraction for the performance characteristics of parallel computers, and should evolve with the development of computational infrastructure. The heterogeneous CPU/Graphics Processing Unit (GPU) systems have been and will be important platforms for scientific computing, which introduces an urgent demand for new parallel computation models targeting this kind of supercomputers. In this research, we propose a parallel computation model called HLog_nGP to abstract the computation and communication features of heterogeneous platforms like TH‐1A. All the substantial parameters of HLog_nGP are in vector form and deal with the new features in GPU clusters. A simplified version HLog₃GP of the proposed model is mapped to a specific GPU cluster and verified with two typical benchmarks. Experimental results show that HLog₃GP outperforms the other two evaluated models and can well model the new particularities of GPU clusters. Copyright © 2015 John Wiley & Sons, Ltd.     

自然计算的仿真研究

针对近年来新兴的计算分支的挑战，如DNA计算、量子计算、免疫计算和进化计算等，提出自然计算的框架、编码、控制策略。以及从自然计算映射到计算机仿真的广义映射模型，来探究新兴计算分支的共同机理及其在自然界中的源泉。借助MATLAB仿真工具对其进行的仿真实验表明，自然计算有优于传统计算的特性，为计算学和计算机科学的发展以及自然计算的仿真开拓了广阔前景和无限生机。     

浅谈高性能计算的地位及应用

高性能计算已被公认为继理论科学和实验科学之后，人类认识世界改造世界的第三大科学研究方法。高性能计算应用在高性能计算技术的支持下为我国的科技创新作出了巨大贡献，并且和高性能计算技术在相辅相成中不断发展。从应用的角度概要总结了高性能计算的作用和地位，列举了几个产业化相关的高性能计算应用，介绍了作为公共服务平台的上海超级计算中心的情况。     

存储与计算的分离

当前计算应用的发展对传统计算机体系结构提出了挑战．由计算资源和存储资源固定连接形成的系统已经不能适应动态计算的需求．从应用出发,提出计算资源和存储资源物理分离和逻辑分离的概念,并以此为基础,构造三维可重构计算环境可以解决这些问题．在这种环境中,用户程序、计算资源和存储资源可以根据应用的需求动态组合．由于摆脱了资源的地理位置和操作环境等方面的限制,计算过程将呈现出数据驱动的特点,从而实现按需计算,使计算系统可以在更大范围内为用户提供服务。