Performance evaluation of ASCI benchmark AZTEC using Paradyn

Performance evaluation of ASCIbenchmark AZTEC using Paradyn Final ReportInstructor:Dr. Frank Mueller (mueller@cs.ncsu.edu)Team Members:Raj kumar Nagarajan (rknagara@cs.ncsu.edu)Vikram S Poojary (vspoojar@cs.ncsu.edu)Problem Description:The main objective of this project is to use Paradyn, a parallelperformance tool, to analyze and evaluate the performance of ASCIbenchmark AZTEC. And to identify the bottlenecks in the benchmarkapplication, to improve the performance by algorithmic changes or MPIspecific changes, suggest improvements to the tool based onexperience with benchmark evaluation. In this document we discuss the following:• AZTEC Benchmark• Performance of AZTEC on the cluster• Paradyn• Evaluation of AZTEC using Paradyn• Improvements to tool• Location of src and binary files on clusterAZTEC:AZTEC is a massively parallel iterative solver library for solving sparselinear systems. It provides state-of-the-art iterative methods thatperforms well on parallel computers (applications of over 200 Gflopshave been achieved on the Sandia-Intel TFlop Computer) and at thesame time is easy to use for application engineers.Simplicity is attained using the notion of a global distributed matrix.The global distributed matrix allows a user to specify pieces (differentrows for different processors) of his application matrix exactly as hewould in the serial setting (i.e. using a global numbering scheme).Issues such as local numbering, ghost variables ...