A Distributed Performance Analysis Architecture for Clusters

The use of a cluster for distributed performance analy-sis of parallel trace data is discussed. We propose an analy-sis architecture that uses multiple cluster nodes as a server to execute analysis operations in parallel and communi-cate to remote clients where performance visualization and user interactions occur. The client-server system devel-oped, VNG, is highly configurable and is shown to perform well for traces of large size, when compared to lead-ing trace visualization systems