Search

The POEMS project is creating an environment for end-to-end performance modeling of complex parallel and distributed systems, spanning the domains of application software, runtime and operating system software, and hardware architecture. Toward this end, the POEMS framework supports composition of component models from these different domains into an end-to-end system model. This composition can be specified using a generalized graph model of a parallel system, together with interface specifications that carry information about component behaviors and evaluation methods. The POEMS Specification Language compiler, under development, will generate an end-to-end system model automatically from such a specification. The components of the target system may be modeled using different modeling paradigms (analysis, simulation, or direct measurement) and may be modeled at various levels of detail. As a result, evaluation of a POEMS end-to-end system model may require a variety of evaluation tools including specialized equation solvers, queuing network solvers, and discrete-event simulators. A single application representation based on static and dynamic task graphs serves as a common workload representation for all these modeling approaches. Sophisticated parallelizing compiler techniques allow this representation to be generated automatically for a given parallel program. POEMS includes a library of predefined analytical and simulation component models of the different domains and a knowledge base that describes performance properties of widely used algorithms. This paper provides an overview of the POEMS methodology and illustrates several of its key components. The methodology and modeling capabilities are demonstrated by predicting the performance of alternative configurations of Sweep3D, a complex benchmark for evaluating wavefront application technologies and high-performance, parallel architectures. [ABSTRACT FROM AUTHOR]

The problem of resource management for many processor architectures can be viewed as the problem of simultaneously updating data structures that hold system state. Consistency requirements of strict data structures introduce serialization in the resource management functions thereby leading to performance bottlenecks in highly parallel systems. Our approach is to examine the possibility of using structures with weakened specifications. Specifically, this paper introduces data structures that weaken the specification of a priority queue permitting it to be updated simultaneously by multiple processes. Two structures are proposed, the concurrent heap and the software banyan along with their associated algorithms for update. The algorithms are shown to possess attractive properties of simultaneous update and throughput. The results of simulation and actual implementations show that such data structures can improve the execution times of parallel algorithms quite significantly. These structures are proposed as possible basic building blocks for implementation of resource allocation in operating systems. [ABSTRACT FROM AUTHOR]

Discusses parallel architectures for computer systems. Usefulness of having several parts of a system simultaneously performing different parts of a task; Potential power in the systems; Models of computation; Languages and architectures; Connecting the elements; Commercial architectures; Specialized processors; Future architectures.