For this ranking we are interested in the time spent

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: loads per cycle (LPC) rates to assess if the code is possibly compute bound. Compile the OpenMP BT solver with a problem size of class A. For a single thread run, determine what are the top six most time consuming routines for this solver. For this ranking, we are interested in the time spent exclusively in the routines, not in any of their callees. By default, the OpenMP run- time will create as many threads as there are cores on your system. But you can use the OMP_NUM_THREADS environment variable to control the number of threads used during a run. Using either PAPI or HPCToolkit, your choice, collect the necessary events to compute the IPC, FPC and LPC metrics for the top six routines identified previously. Once again,...
View Full Document

This document was uploaded on 03/15/2014 for the course COSC 594 at University of Tennessee.

Ask a homework question - tutors are online