The successful application of optimistic synchronization techniques in parallel simulation requires that rollback overheads be contained. The chief contributions to rollback overhead in a Time Warp simulation are the time required to save state information and the time required to restore a previous state. Two competing techniques for reducing rollback overhead are periodic checkpointing (Lin and Lazowska, 1989) and incremental state saving (Bauer et al., 1991). This paper analytically compares the relative performance of periodic checkpointing to incremental state savings. The analytical model derived for periodic checkpointing is based almost entirely on the previous model developed by Lin (Lin and Lazowska, 1989). The analytical model for incremental state saving has been developed for this study. The comparison assumes an optimal checkpoint interval and shows under what simulation parameters each technique performs best.
This paper addresses one problem of speeding parallel digital system simulation using Time Warp. In particular, one problem with Time Warp is that logical processes with errant behavior can incur considerable rollback behavior (analogous to thrashing in paging virtual memories). Consequently, additional mechanisms must be added to an optimistically synchronized simulator to inhibit excessive rollback. Herein, we describe a method of adaptively sizing bounded time windows to balance lookahead processing.
Optimistic techniques using the Time Warp mechanism has shown great promise in speeding up Parallel Discrete Event Simulations. However, Time Warp has been plagued by problems such as excessive rollbacks, memory usage, and wasted lookahead computation. In particular, excessive rollbacks can result in a deterioration of the advancement of the simulation. Consequently, techniques to properly control the optimism in Time Warp are needed to alleviate nonproductive lookahead. This paper presents a Logical Process (LP) scheduling algorithm based on concepts from adaptive control theory. In particular, we develop a performance index called useful work. The useful work parameter represents the amount of productive work done by the process and it is used by our scheduling algorithm to aid in LP scheduling decisions. The scheduling algorithm presented in this paper is compared with the widely used smallest timestamp first scheduling algorithm to show its usefulness in a Time Warp simulation. Palaniswamy, A. and Wilsey, P. A. (199313). 765-770. SCS. AUTHOR BIOGRAPHIES AVINASH C. PALANISWAMY completed his Ph.D. in Electrical and Computer Engineering from the University of Cincinnati in May, 1994. His research interests include, parallel synchronization algorithms, high speed parallel simulation of digital systems, CAD for board level simulation, Hardware/Software Co-Simulation, and acceleration of VHDL simulation. PHILIP A. WILSEY is an assistant professor at the University of Cincinnati. His primary research interests are computer architecture, parallel processing, hardware description languages, MIMD on SIMD, and CAD.
Technological advances have resulted in the design of large integrated circuits, and with it the need for fast simulation to decrease the design-to-market time. Parallel and distributed digital simulation has been recognized as the leading approach to provide fast simulation of digital circuits. The Time Warp Mechanism (T W M ) implement:? the concept of Virtual Time t o optimistically synchronize parallel simulation. However, the successful utili.ration of the Time Warp mechanism has been plagued by the time and space overheads of rollback, namely: state saving, state restoration, and event reprocessing. Several optimizations to reduce these overheads have been proposed in the literature. This paper presents empirical data relating the effectiveness of several of these Optimizations to the domain of digital system simulation. In particular, we present performcince results from lazy reevaluation, lazy cancellation, periodic state savings, and bounded time windows.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.