If the HP Caliper run is made on a ccNUMA system, then the memory usage of every “logical domain” is separately measured and reported. If on an SMP system, then only the single, “local domain” is measured and reported.
The system memory usage measurement is always taken if the
The process memory usage measurement can be taken:
•At the beginning of a process's execution:
•At the end of a process's execution:
•Periodically throughout a process's execution:
•At any combination or all of the above:
The default sampling rate for measuring process memory usage is 1 second. To specify a different rate, use an integer value (greater than 0) followed by an optional time modifier indicating seconds (s), minutes (m), or hours (h). The value is assumed to be in seconds if no time modifier is given.
To reduce the volume of data, only samples that show a difference in memory utilization from the previous sample are saved and reported.
NOTE: To maintain compatibility with earlier versions of HP Caliper, exit is synonymous with end.
The process memory usage being measured is the number of pages the process has allocated from the system. It includes executable code, shared libraries, static data, mmap'd regions, and dynamic (malloc'd heap) memory. It does not measure how much of this
The
As with all HP Caliper
Examples of the
Some examples of the option follow:
•
Causes process memory usage to be measured at the beginning, at the end, and every 1 second of the process's execution.
•
Causes process memory usage to be measured twice: at the beginning and at the end of the process's execution.
•
Causes process memory usage to be measured every 15 seconds of the process's execution.
•
Causes process memory usage to be measured at the beginning, at the end, and every 30 minutes of the process's execution.
•
Causes process memory usage to be measured once: at the end of the process's execution.
Measuring Memory Usage Concurrently with Other Measurements 153