This paper studies the scalability issues of Java Virtual Machine (JVM) on Symmetrical Multiprocessing (SMP) systems. Using a cycle-accurate simulator, we evaluate the performance scaling of multithreaded Java benchmarks with the number of processors and application threads. By correlating low-level hardware performance data to two high-level software constructs: thread types and memory regions, we present in detail the performance analysis and study the potential performance impacts of memory system latencies and resource contentions on scalability.Several key findings are revealed through this study. First, among the memory access latency components, the primary portion of memory stalls are produced by L2 cache misses and cache-to-cache transfers. Second, among the regions of memory, Java heap space produces most memory stalls. Additionally, a large majority of memory stalls occur in application threads, as opposed to other JVM threads. Furthermore, we find that increasing the number of processors or application threads, independently of each other, leads to increases in L2 cache miss ratio and cache-to-cache transfer ratio. This problem can be alleviated by using a thread-local heap or allocation buffer which can improve L2 cache performance. For certain benchmarks such as Raytracer, their cache-to-cache transfers, mainly dominated by false sharing, can be significantly reduced. Our experiments also show that a thread-local allocation buffer with a size between 16KB and 256KB often leads to optimal performance. 119 0-7803-9461-5/05/$20.00 ©2005 IEEE

System-level design (SLD) provides a solution to the challenge of increasing design complexity and time-to-market pressure in modern embedded system designs. In this paper, we propose a novel system-level approach to communication architecture modeling, which was not yet well addressed in existing SLD methodologies. In particular, we show how to develop statistical models for communication architectures. These new models are capable of capturing communication details at higher abstraction levels than previously possible. We demonstrate how to use the mean-square-error as a tool for developing these models, and show where to integrate these models in the design process.

No abstract