Livelocks in parallel programs

This paper is the second of a two-part series exploring the subtle correctness criterion of the absence of livelocks in parallel programs. In this paper we are concerned with the issue of proving this correctness criterion. It is shown that livelocks are not preserved by reduction, implying that reduction cannot be used directly in proving the absence of livelocks. Two applicable proof techniques are also presented. One is based on the notion of establishing sufficient conditions for livelock-freedom; the other is an extension of the well-founded set method for proving termination in sequential programs.     

信息转接死锁避免的动态优化算法

对于动态传输路径，给出了每个结点至少两个缓冲的信息转接网死锁避免问题，所提出的死锁避免算法从缓冲利用的角度看是优化的。同时给出了死锁无关的证明及其条件，这对于并行计算处理问题是有益的。     

基于未来锁集的死锁规避

针对现有动态死锁规避方法存在能力有限、被动盲目、开销较大和影响目标程序正确性等问题,提出一种基于未来锁集的动静结合死锁规避方案Flider.基本思想是,对于一个加锁操作,若其未来锁集中的所有锁都是空闲的,则执行该加锁操作不会导致死锁.一个加锁操作的未来锁集包括当前要加锁的锁和从该加锁操作到与之相对应的解锁操作过程中遇到的所有加锁操作所要加的锁.通过静态分析,计算锁效应信息并插桩到相应的加锁操作和函数调用操作前后.通过动态分析,劫持加锁操作,根据其锁效应信息为之计算未来锁集,只有当未来锁集中的所有锁都未被锁定才执行该加锁操作,否则等待.测评实验和对比实验表明Flider能智能主动地规避多种类型死锁,开销较小,扩展性好,不影响程序正确性.     

Extension of the lower bound of monitor solutions of maximally permissive supervisors to non-α net systems

Traditional region-based liveness-enforcing supervisors focus on (1) maximal permissiveness of not losing legal states, (2) structural simplicity of minimal number of monitors, and (3) fast computation. Lately, a number of similar approaches can achieve minimal configuration using efficient linear programming. However, it is unclear as to the relationship between the minimal configuration and the net structure. It is important to explore the structures involved for the fewest monitors required. Once the lower bound is achieved, further iteration to merge (or reduce the number of) monitors is not necessary. The minimal strongly connected resource subnet (i.e., all places are resources) that contains the set of resource places in a basic siphon is an elementary circuit. Earlier, we showed that the number of monitors required for liveness-enforcing and maximal permissiveness equals that of basic siphons for a subclass of Petri nets modelling manufacturing, called α systems. This paper extends this to systems more powerful than the α one so that the number of monitors in a minimal configuration remains to be lower bounded by that of basic siphons. This paper develops the theory behind and shows examples.     

Parallel simulation on the hypercube multiprocessor

Summary This paper focuses upon a particular conservative algorithm for parallel simulation, the Time of Next Event (TNE) suite of algorithms [13]. TNE relies upon a shortest path algorithm which is independently executed on each processor in order to unblock LPs in the processor and to increase the parallelism of the simulation. TNE differs fundamentally from other conservative approaches in that it takes advantage of having several LPs assigned to each processor, and does not rely upon message passing to provide lookahead. Instead, it relies upon a shortest path algorithm executed independently in each processor. A deadlock resolution algorithm is employed for interprocessor deadlocks. We describe an empirical investigation of the performance of TNE on the iPSC/i860 hypercube multiprocessor. Several factors which play an important role in TNE's behavior are identified, and the speedup relative to a fast uniprocessor-based event list algorithm is reported. Our results indicate that TNE yields good speedups and out-performs an optimized version of the Chandy&Misra-null message (CMB) algorithm. TNE was 2–5 times as fast as the CM approach for less than 10 processors (and 1.5–3 times as fast when more than 10 processors were used for the same population of processes.) Azzedine Boukerche received the State Engineer degree in Software Engineering from Oran University, Oran, Algeria, and the M.Sc. degree in Computer Science from McGill University, Montreal, Canada. He is a Ph.D. candidate at the School of Computer Science, McGill University. During 1991–1992, he was a visiting doctoral student at the California Institute of Technology. He is employed as a Faculty Lecturer of computer Science at McGill University since 1993. His research interests include parallel simulation, distributed algorithms, and system performance analysis. He is a student member of the IEEE and ACM. Carl Tropper is an Associate Professor of Computer Science at McGill University. His primary area of research is parallel discrete event simulation. His general area of interest is in parallel computing and distributed algorithms in particular. Previously, he has done research in the performance modeling of computer networks, having written a book,Local Computer Network Technologies, while active in the area. Before coming to university life, he worked for the BBN Corporation and the Mitre Corporation, both located in the Boston area. He spent the 1991–92 academic year on a sabbatical leave at the Jet Propulsion Laboratories of the California Institute of Technology where he contributed to a project centered about the verification of flight control software. As part of this project he developed algorithms for the parallel simulation of communicating finite state machines. During winters he may be found hurtling down mountains on skis.This work has been completed while the author was a visiting doctoral student at the California Institute of TechnologyWas on sabbatical leave at the Jet Propulsion laboratories, California Institute of Technology     

Controllability of control and mixture weakly dependent siphons in S3PR

Deadlocks in a flexible manufacturing system modelled by Petri nets arise from insufficiently marked siphons. Monitors are added to control these siphons to avoid deadlocks rendering the system too complicated since the total number of monitors grows exponentially. Li and Zhou propose to add monitors only to elementary siphons while controlling the other (strongly or weakly) dependent siphons by adjusting control depth variables. To avoid generating new siphons, the control arcs are ended at source transitions of process nets. This disturbs the original model more and hence loses more live states. Negative terms in the controllability make the control policy for weakly dependent siphons rather conservative. We studied earlier on the controllability of strongly dependent siphons and proposed to add monitors in the order of basic, compound, control, partial mixture and full mixture (strongly dependent) siphons to reduce the number of mixed integer programming iterations and redundant monitors. This article further investigates the controllability of siphons derived from weakly 2-compound siphons. We discover that the controllability for weakly and strongly compound siphons is similar. It no longer holds for control and mixture siphons. Some control and mixture siphons, derived from strongly 2-compound siphons are not redundant – no longer so for those derived from weakly 2-compound siphons; that is all control and mixture siphons are redundant. They do not need to be the conservative one as proposed by Li and Zhou. Thus, we can adopt the maximally permissive control policy even though new siphons are generated.     

Solving siphons with the minimal cardinality in Petri nets and its applications to deadlock control

Siphons can be used to characterise deadlock states and solve deadlock problems in Petri nets that model flexible manufacturing systems. By modifying the objective function and adding new constraints to the mixed-integer programming (MIP) method proposed by Park and Reveliotis, this paper presents a revised MIP (RMIP) to directly solve siphons, called smart siphons, with the minimal cardinality as well as the minimal number of resource places. Accordingly, an iterative siphon-based control (ISC) method adds a proper control place (CP) to make each smart siphon max-controlled until the controlled system is live. Since any siphon that has more resource places can be composed of those containing less resource places, a small number of CPs are required during the iterative control process due to the solved smart siphons containing the minimal number of resource places. Compared with the existing methods in the literature, the proposed ISC method using the RMIP avoids computing a maximal deadly marked siphon from which a minimal siphon is then derived, adds a small number of proper CPs, and leads to a liveness-enforcing supervisor with a simple structure. Finally, a case study demonstrates the effectiveness of the proposed ISC method.     

Locked discrete event systems: How to model and how to unlock

To model qualitative aspects of discrete event systems, i.e., the order of the events is of sole importance, we use a triple consisting of the set of all possible events (the alphabet), the set of all behavior (possible strings of events), and the set of all tasks (completed behavior). We use this view to model synchronous as well as asynchronous connection of systems. Moreover, it is easy to define notions like deadlock and livelock in this view. We give a method to construct a second system that, in connection with the original system, gets rid of its deadlock and/or livelock. A state-space representation is introduced. In this representation computations can be done effectively.