Samir R. Das, Richard M. Fujimoto
The performance of the Time Warp mechanism is experimentally evaluated when only a limited amount of memory is available to the parallel computation. An implementation of the cancelback protocol is used for memory management on a shared memory architecture to assess the performance versus memory tradeoff. Through various synthetic and benchmark programs, the study investigates how limited memory influences Time Warp behavior. Key findings indicate that allowing for the cancellation of multiple memory objects can significantly improve performance, particularly when memory is constrained. Additionally, the research reveals that there exists a well-defined speedup versus memory curve, showing rapid performance benefits with increased memory up to a certain limit, beyond which gains are minimal. The study concludes that with effective memory management, performance near that of full-memory operation can be achieved with only a modest increase in memory, challenging the notion that significant memory is essential for optimal Time Warp execution.
@article{d8b5fcf1-5e4c-4609-a0e4-f56268ef3f77,
title={An empirical evaluation of performance m},
author={Samir R. Das and Richard M. Fujimoto},
year={2026},
language={en}
}TY - JOUR TI - An empirical evaluation of performance m AU - Samir R. Das AU - Richard M. Fujimoto PY - 2026 LA - en ER -
This paper addresses the challenge of assessing the feasibility of wind power plant projects at sites with insufficient or no local historic wind data
Important advances in electrochemical engineering technology over the last three decades have fostered the development of a lternative methods to alle
Increasing volumes of waste printed circuit boards from obsolete electronic equipment posed escalating environmental risks and resource losses due to
The leachability tests for manufacturing scrap TV boards (STVB) have indicated the release of metals beyond the limit levels with potential problems f