Parallel Performance Project (PPP)
Contributed Shared Memory Tools

The Parallel Performance Project in the University of Michigan is making a suite of its research tools available to other researchers. These tools are intended to help in developing better shared memory applications and systems. They enable trace collection, application analysis, existing-system calibration, and system simulation.

This suite includes the following tools:

• The Shared-Memory Application Instrumentation Tool (SMAIT)
  
  
• The Configuration Independent Analysis Tool (CIAT)
  [NOTE: Special usage restrictions apply to the CIAT tool (see license agreement below).]
  
  
• The Time Distribution Analysis Tool (TDAT)
  
  
• Microbenchmarks for calibrating a distributed shared memory system (DSM-MB)
  
  
• The Communication Contention Analysis Tool (CCAT) -- Will be added later --
  
  

License Agreement

The source code of these tools is provided through this web page by the Hewlett-Packard Company, Copyright (c) 1996, and the University of Michigan, Copyright (c) 1997-1998. All rights reserved.

Individuals affiliated with NPACI, government, or academia are granted permission to download, use, and modify these tools, without fee, provided that they acknowledge the source of these tools in any publication of research that uses these tools, and they do not use these tools for commercial purposes. Individuals affiliated with commercial companies are not granted the right to download or use CIAT.

By downloading these tools, you accept the terms of this license.

Disclaimer

Credits

• These tools were developed primarily by Gheith Abandah of the PPP research group under the supervision of Professor Edward Davidson.
  
  
• The development of SMAIT, CIAT, and TDAT started in 1996 at Hewlett-Packard Laboratories, Palo Alto, California. These tools were further enhanced and extended at the University of Michigan in 1997 and 1998.
  
  
• SMAIT is based on a tool called RYO developed by Daniel Zucker and Alan Karp.
  
  
• DSM-MB were developed in the University of Michigan, and were used to calibrate the Convex SPP1000 in the Center for Parallel Computing at the University of Michigan (CPC) and the HP/Convex SPP2000 in the Center for Advanced Computing Research at Caltech (CACR).
  
  
• Tomas Rokicki from HP Labs has contributed software modules that are used in CIAT and CCAT.
  
  
• CCAT uses the same front-end trace reading routines that are used in CIAT.
  
  
• CCAT uses event scheduling routines that are adapted from the SimPack toolkit developed by Paul A. Fishwick.
  
  

Click here to learn more about these tools

Click here to download

Last updated: August 20, 1998.