This section describes the procedures and options for collecting performance data for stand-alone (client) applications by using the Visual Studio Profiling Tools from the command line.
Collect application statistics: Use the sampling method to collect performance statistics. Sampling data is useful for analyzing CPU utilization issues and for understanding the general performance characteristics of an application.
Collect detailed timing data: Use the instrumentation method to collect detailed timing information. Instrumentation data is useful for analyzing I/O issues and for fine-grained analysis of application scenarios.
Collect .NET memory data: Use sampling or instrumentation to collect .NET memory allocation data that shows you the size and number of allocated objects. You can also collect object lifetime data that shows you the size and number of objects that are reclaimed in each garbage collection generation.
Collect concurrency data: Use the concurrency method to collect resource contention data and thread activity data that shows you CPU utilization, thread contention, thread migration, synchronization delays, areas of overlapped I/O, and other system events.
Add tier-interaction data: You can add performance data about synchronous ADO.NET calls that the application made to a Microsoft SQL Server database. Adding tier interaction data to a profiling run requires specific procedures with the command line profiling tools.