This topic describes two fundamental types of program execution - data parallelism and task parallelism - and the task patterns of each.
You need to specify the directory containing the Intel Advisor definition file as an additional include directory when you compile your program. Intel Advisor installs its annotation definition files into a default directory on your system. For example:
Use this special-purpose annotation to stop tracking references to a memory location by the Dependencies tool. This information can help you understand what code accesses a memory location. When you have learned enough, simply remove this annotation.
For medium-large targets, several methods are available to minimize the amount of data collected and target execution time. Minimizing the data collected reduces the amount of data you need to examine in the Suitability Report; it also reduces the size of the generated result.
Lock overhead is the time spent in creating, destroying, acquiring, and releasing locks. Lock overhead does not include the time spent waiting for a lock held by another task - that is called lock contention. You can think of lock overhead as the cost of the lock operations themselves assuming the lock is always available.
Data collected by running Intel® Advisor tools is stored in a result. When you run one of its tools, the Intel® Advisor executes a target, identifies issues that may need handling, collects the results and shows it in the Results subdirectory in the Solution Explorer in Microsoft Visual Studio* or in the Project Navigator in the Intel AdvisorStandalone GUI.
When running Intel Advisor Dependencies tool analysis, you see unmatched problems reported that are caused by unmatched annotations execution that you did not expect.
Enable task chunking for specified annotated sites when generating a suitability report.
Examine specified sites for opportunities to reduce task overhead.
This tab provides assembly representation of the currently selected code site.