What If Home | Product Overview | Features | Technical Requirements | Documentation FAQ | Primary Technology Contacts

Product Overview

Intel Software Autotuning Tool

The Intel Software Autotuning Tool (ISAT) is a new performance tuning tool that supports automatic searching for the near-optimal values of program parameters that have significant performance impact. Examples are the cache-blocking factors in matrix computations, the task granularity in TBB, and the scheduling policy of OpenMP parallel constructs. The output of ISAT is a tuned version of the program based on the searching results. In addition, it supports visualization of the tuning results (e.g., Figure 1) so that the developer can have a better understanding of his program performance. ISAT can be used on top of any C/C++ compilers on Linux.

Figure 1: Visualization of ISAT results.

Features

• Automatic searching of the near-optimal values of program parameters
• Easy to use with no compiler dependency
• Programmer-controlled searching strategies
• Threading-library specific tuning for OpenMP and TBB
• Visualization of tuning results
• The autotuner is open-sourced

Technical Requirements

• 32 or 64-bit Linux; Python version between 2.4 and 2.9; any C/C++ compiler.

Downloads

• Kit for 32/64-bit Linux

Documentation

• User manual
• A Synergetic Approach to Throughput Computing on x86- Based Multicore Desktops
• Presentation: A Synergetic Approach to Throughput Computing on Intel® Architecture

Primary Technology Contacts

Chi-Keung (CK) Luk is a Senior Staff Engineer in a Software Pathfinding Group at Intel, where he conducts research and advanced development in parallel programming, compiler, and program-analysis tools. Previously, he worked on the Pin dynamic instrumentation system. CK obtained his Ph.D. from the University of Toronto and was a visiting scholar at Carnegie Mellon University. He has 30 publications and two issued patents with a few others pending. He received an Intel Achievement Award and a nomination for the ACM Doctoral Dissertation Award.