Program Section Usage Summary report

The Program Section Usage Summary report shows the distribution of CPU time used by each active control section of each module in the target program or subsystem. When you create the Performance Profile and specify the ALLCSECT parameter, the report also displays inactive control sections. Examine this report to determine what programs or systems services are consuming the most CPU resources. You can then look up those programs in the Program Usage by Procedure report to determine which procedures cause the consumption.

Program Section Usage Summary Report

CPU time used by each control section is given as a percentage of all CPU time used in the measured job step; thus, the CPU time percentages used by all listed control sections total 100. The report orders entries alphabetically by module name and then section name.

The report groups all system service routines by function under appropriately named pseudo-sections within the pseudo-module .SYSTEM. It groups compiler library routines included in the target program or subsystem load modules under a suitably named pseudo-section.

MODULE NAME

The name of the module or pseudo-module for which Strobe is detailing CPU usage. For module names that exceed eight characters, see the note in the SECTION NAME paragraph.

SECTION NAME

The name of the control section or the pseudo-section within the module or pseudo-module identified in MODULE NAME for which Strobe is detailing CPU usage. A section name may not appear if the activity in the section is less than the percentage specified in the MODULE MAPPING BASELINE (default 2%) or if Strobe was unable to map the module. Strobe is unable to index a section that is not mapped.

Note

To ensure that the control section information is complete, you must compile all VisualAge products with the binder option EDIT=YES (the default).

For all module or section names that exceed eight characters, Strobe generates a token, which is an eight-byte identifier. The token comprises the first four characters of the module or section name followed by a hyphen (-) and then the last three characters of the name. See the Token - Longname Cross Reference report to reconcile all tokens with their long names.

Note

Strobereports ?LONGnnn for the section names that exceed eight characters if initialization of the long names data space fails. The suffix nnn is a sequence number beginning with 001. The suffix is reset to 001 for each module. If initialization of the long names data space fails, Strobe also does not produce the Token - Longname Cross Reference report.

16M < , >

Indicator whether the RMODE of a module is 24 bit (<) or 31 bit (>). If the module has multiple control sections, this indication appears next to TOTALS in the SECTION NAME column. If the module is a Generalized Object File Format (GOFF) split-RMODE module an S appears in this column. If a GOFF module contains multiple control sections, the RMODE indicator appears for each control section.

SECT SIZE

The size of the control section in bytes, expressed in decimal notation. The report does not show section size for pseudo-sections, such as .IOCS and .SVC, that group related system services together and summarize those services.

FUNCTION

A short description of the function of the control section or pseudo-section, or of the module or pseudo-module. Function descriptors appear for all pseudo-sections and pseudo-modules and, if your Strobe system programmer has supplied them, for other control sections and modules as well.

CPU TIME PERCENT

The percentage of all CPU time executing on behalf of the control section. There are two measures of CPU time:

- SOLO shows activity without any concurrent I/O activity for the target program or subsystem.
- TOTAL shows activity with or without concurrent I/O activity.

If a module contains more than one control section, the report also shows subtotals by control section. The Program Usage by Procedure report treats each control section in the Program Section Usage Summary report in greater detail.

CPU TIME HISTOGRAM

The intensity of CPU usage within each control section. Solo CPU time is indicated by the asterisk symbol (*). The remaining CPU time is indicated by the plus symbol (+). Spikes, or lengthy lines in the histogram, highlight control sections with a high proportion of activity.

MARGIN OF ERROR

The margin of error for the CPU time percentages, also reported on the Measurement Session Data report. This margin of error, which appears in the header line, applies only to the number of samples in which Strobe found the CPU to be active.