Quantifying the impact of dynamic memory managers into memory-intensive applications

Abstract

Modern portable devices execute multimedia applications that exhibit high resource utilization. To efficiently execute these applications in embedded systems, the dynamic memory subsystem needs to be optimized. This complex task can be tackled designing custom dynamic memory management mechanisms. Currently, several automatic methodologies to optimize custom Dynamic Memory Managers (DMMs) have been proposed. However, these approaches are mainly related to improve application performance. In this paper we propose a methodology to automatically evaluate the impact of any DMM into an application considering four different metrics: performance, memory usage, temperature and energy consumption. This methodology is applied to Lea, a well-known general-purpose memory allocator. Our experimental results over five different memory-intensive applications show that, on average, Lea consumes a 43.25% and 22.90% of execution time and memory usage, respectively. In addition, the memory temperature and energy consumed, related only to the memory device, are increased by 0.39% and 0.48%, respectively.

J. Manuel Colmenar

Associate Professor

My research interests are focused on metaheuristics applied to optimization problems. I have worked on different combinatorial optimization problems applying trajectorial algorithms such us GRASP or VNS. Besides, I am very interested in applications of Grammatical Evolution, specifically in model and prediction domain, as alternative to machine learning approaches.