Choosing a Small, Representative Data Set

When you run an analysis, the Intel Advisor executes the target against the supplied data set. Data set size and workload have a direct impact on application execution time and analysis speed

For example, it takes longer to process a 1000x1000 pixel image than a 100x100 pixel image. A possible reason: You may have loops with an iteration space of 1...1000 for the larger image, but only 1...100 for the smaller image. The exact same code paths may be executed in both cases. The difference is the number of times these code paths are repeated.

You can control analysis cost without sacrificing completeness by minimizing this kind of unnecessary repetition from your target's execution.

Instead of choosing large, repetitive data sets, choose small, representative data sets that fully create tasks with minimal to moderate work per task. Minimal to moderate means just enough work to demonstrate all the different behaviors a task can perform.

Your objective: In as short a runtime period as possible, execute as many paths and the maximum number of tasks (parallel activities) as you can afford, while minimizing the repetitive computation within each task to the bare minimum needed for good code coverage.

Data sets that run in about ten seconds or less are ideal. You can always create additional data sets to ensure all your code is checked.

To modify the input data set in the Visual Studio IDE, do one of the following:

  • Specify Properties for the project or configuration. For example, right-click the startup project's name to display the context menu:
    1. Choose Properties > Configuration properties > Debugging.
    2. Select the type of configuration this change will apply to by selecting the type under Configuration, such as Active(Debug), Debug, Release, or All Configurations. By default, properties for Debug and Release configuration are maintained separately.
    3. Edit the Command Arguments to select the appropriate data set.
    4. Click OK.
  • Specify a different startup project that already has a reduced data set.
  • Modify the program's sources (perhaps using #ifdef directives) and rebuild the target.

To modify the input data set using the Intel Advisor GUI, do one of the following

  • Specify the project properties for the target. For example:
    1. Either click File > Project properties... or the icon on the Intel Advisor toolbar. This displays the Project Properties dialog box.

    2. If needed, click the Analysis Target tab.

    3. In the Target type drop-down list, choose Dependencies Analysis.

    4. In the Application parameters, if your target's main entry point accepts command-line arguments, specify a value in this field. Either type a value, or click the Modify... button.

    5. When done, click OK.

  • Modify the program's sources (perhaps using #ifdef directives) and rebuild the target.

Tip

  • If you run this configuration often, consider creating a new configuration perhaps called Dependencies for this small data set.

  • For the most current information on optimal C/C++ and Fortran build settings, see Getting Started with Intel® Advisor.

Tip

For the most current information on optimal C/C++ and Fortran build settings, see Getting Started with Intel® Advisor.

For more complete information about compiler optimizations, see our Optimization Notice.