Windows System Resource Manager samples

The Windows System Resource Manager CD-ROM includes samples that illustrate the features and capabilities of Windows System Resource Manager. You can use these samples to evaluate the performance of your system when Windows System Resource Manager management is enabled.

The samples include of the following:

Sample file Description
Allocationpol.xml Includes the following resource-allocation policies:
  • Equal_Per_User - The policy that is included with Windows System Resource Manager. It implements the equal-per-user management rule and specifies a 99% CPU consumption target.
  • DayTimePolicy - This policy is designed to manage two of the sample programs, Batch.exe and Bizcrit.exe. For Batch.exe, a CPU consumption target of 70% is specified. For Bizcrit.exe, a CPU consumption target of 25% is specified.
  • NightTimePolicy - This policy is designed to manage two of the sample programs, Batch.exe and Bizcrit.exe. For Batch.exe, a CPU consumption target of 60% is specified. For Bizcrit.exe, a CPU consumption target of 30% is specified.
  • StopLeak - This policy is designed to end a program that has a memory leak when its memory consumption reaches a specified level.
Calendar.xml Activates the default post-installation calendar settings. These settings are: calendar management, the calendar default policy set to WSRMDefault, and the managing policy set to WSRMDefault.
Selectionpol.xml Includes the following process-matching criteria:
  • Equal_Per_User - The process-matching criterion that is included with Windows System Resource Manager. It matches all running processes not included on either the system-defined exclusion list or the user-defined exclusion list.
  • Batch - Matches only Batch.exe
  • BizCrit - Matches only Bizcrit.exe
  • Rogue - Matches only Rogue.exe
  • Leaky - Matches only Leaky.exe
  • IIS-Default - Matches any W3wp.exe process that is serving Web pages out of the Internet Information Services (IIS) 6.0 default application pool.
Batch.exe Represents a non-interactive workload that consumes more resources when interactive workloads consume fewer resources.
Bizcrit.exe Represents an interactive workload that a business might depend on, such as customer-service software or a Web site.
Rogue.exe Represents a CPU-intensive workload that has a low business priority, but which must still run, provided that it does not impact other running programs that are more important.
Leaky.exe Represents a program with a memory leak.

Note

Using the samples

You can use the samples to demonstrate and compare the differences between the CPU consumption of an application when the computer is managed by Windows System Resource Manager and when it is not managed by Windows System Resource Manager.

On your Windows System Resource Manager CD-ROM, go to the Samples directory. Open the x86 folder, or, if you are using an Intel Itanium-based processor, open the ia64 folder. You can either copy the four sample programs (Bizcrit.exe, Batch.exe, Rogue.exe, and Leaky.exe) to any directory on your local computer, or run the sample programs off the CD-ROM. You can run the sample programs from any location. To run the samples, double-click one or more of the sample programs. The sample programs will start. An empty console application window appears for each program you start.

When the samples are running, go to the Administrative Tools folder, and then double-click Performance to start System Monitor. In System Monitor, select the Process performance object, and then select % Processor Time for the each of the running sample processes (Batch, Bizcrit, Leaky, or Rogue). System Monitor will display the CPU consumption for each process. You can use this information to measure the performance of the system.

After you have started one or more sample programs, you can compare the performance of the system in the managed state to its performance in the unmanaged state. To do so, open the Windows System Resource Manager snap-in. Enable management and use System Monitor to measure the performance of the system. Then disable management and use System Monitor to measure the performance of the system in the unmanaged state. You can then make your comparison.