Development of a Parallel Programming Environment with Dynamic Resource Management Facilities

我们设计并实现了一个新的并行程序设计环境Parsley，它提供基于应用程序结构的细粒度调度服务。在Parsley中，应用程序被划分为可以串行或并行运行的子任务。它提供了一个编程接口，允许用户定义子任务并轻松指定它们之间的优先约束。Parsley使用这些约束在运行时调度子任务。日程安排政策。自动改进以反映硬件环境和资源使用情况。基本调度策略是基于关键路径法的增量式调度算法。在该算法中，子任务的优先级是动态确定的，通过使用每个子任务的执行时间，由资源管理设施的Parsley。我们在Parsley系统上开发了一个并行分子动力学（MD）模拟程序，并在三个可扩展的多处理器系统上执行。我们发现，Parsley是有效的大规模MD模拟和负载平衡设施的Parsley可以采用不同架构的可扩展的多处理器系统。Parsley上的MD模拟比传统的并行算法快3.5倍。Parsley对调度策略的自动改进进一步提高了35%到55%的处理器利用率。此外，Parsley在异构环境（例如不同工作站和集群的网络）以及同构环境中都很有用。由于Parsley的动态处理器分配功能，用户不需要知道计算机和网络的单独性能。

We designed and implemented a new parallel programming environment called Parsley, which provides fine-grained scheduling services based on the structures of application programs. In Parsley, application programs are divided into subtasks that can run serially or in parallel. It provides a programming interface that allows a user to define subtasks and to easily specify precedence constraints among them. Parsley uses these constraints to schedule subtasks at run time. The scheduling' policy. is automatically improved to reflect the hardware environment and resource usage. The basic 'scheduling policy is an incremental scheduling algorithm based on the critical path method. In this algorithm, subtask priorities are dynamically determined by using the execution time of each subtask, as monitored by the resource management facilities of Parsley. We developed a parallel molecular dynamics (MD) simulation program on the Parsley system and' executed it on three scalable multiprocessor systems. We found that Parsley is efficient for large-scale MD simulation and that load balancing facilities of Parsley can be adopted to the scalable multiprocessor systems of different architectures. The MD simulation on Parsley is 3.5 times faster than the conventional parallel algorithm. The Parsley's automatic improvement of scheduling policies further enhances processor utilization by 35 to 55 %. In addition, Parsley is useful in the heterogeneous environment (e.g. a network of different workstations and clusters) as well as the homogeneous environment. Users need not be aware of the individual performance of the computers and networks, because of the dynamic processor allocation facilities of Parsley.