The performance-measurement compiler option ENBPFRCOL() allows you to specify whether or not the compiler should generate code (sometimes called performance hooks) into your compiled program or module. The performance hooks enable the Performance Explorer to analyze your programs. The default for this option specifies that program entry procedure level performance-measurement code is generated for your compiled module or program.
Compiling performance collection code into the module or program allows performance data to be gathered and analyzed. The insertion of the additional collection code results in slightly larger module or program objects and may affect performance to a small degree.
Types of performance data collected include:
This information is gathered immediately before and after calling any given functions. It provides a record of where a call was made, and information on the performance of the operation called.
This information is gathered immediately upon entry into a procedure and exit from that procedure. A snapshot is taken of the current performance statistics when entering a procedure, and a calculation is made of the differences in those statistics when exiting that procedure.
When performance collection code is generated into a leaf procedure, the procedure is changed so that it is no longer a leaf procedure. (A leaf procedure is one that does not call any other procedures.) This is because the leaf procedure now contains hooks to call the performance collection routines. This can be a time-consuming process.
See WebSphere Development Studio: ILE C/C++ Compiler Reference for information on these options.
(C) Copyright IBM Corporation 1992, 2005. All Rights Reserved.