Runtime Analysis in Release 4.6A
1 Measurement in parallel session
2 Measurement restrictions
3 Aggregation
4 Measurement variants
5 Unique identification of internal tables
6 Measurement for particular statements
7 Hit lists
8 Use of new tools
9 Changes to the user interface
ABAP_MODIFICATION_1 Measurement in Parallel Session
You can activate runtime analysis for ten simultaneous external sessions
(work processes). This makes it possible to switch into a long-runnng
job from outside. You can switch in and out of the work processes that
you want to measure using a screen similar to Transaction SM50 .
ABAP_MODIFICATION_2 Measurement Restrictions
You can restrict the duration of a measurement, the size of the
performance data file, and also select the objects that you want to
measure.
The duration of the measurement is specified in minutes.
The measurement ends after this time. The maximum duration is
approximately 72 minutes.
The size of the performance data file is specified in kilobytes. Once
this size is reached, the measurement ends.
You can restrict the measurement to specific objects:
Programs and function groups
Global and local classes
Methods, subroutines, or function modules
Particular statements
ABAP_MODIFICATION_3 Aggregation
Aggregation can be divided into full aggregation and
aggregation by call position. In full aggregation , all
performance data from the same event is summarized in a single record.
In aggregation by call position, all performance data from a particular
event onwards is summarized.
This allows a multiple-level analysis in which the first step is to
determine the most runtime-intensive statements. Then, in a second step,
you can run a new measurement restricted to these objects.
Some information is missing from the aggregation: There is no
hierarchy. Furthermore, you cannot branch to the call location from
full aggregation . Due to the display filter, the new results are
displayed wtihout the net times being recalculated.
ABAP_MODIFICATION_4 Measurement Variants
The default variant is stored under the name DEFAULT . If you
change this variant, the changed data is stored temporarily under the
name USER_DEFAULT . If you find that you are using the same
restrictions frequently, it is worth creating a variant. A variant
exists only in the system in which you created it, and it can be used,
changed, or deleted by any user.
ABAP_MODIFICATION_5 Unique Identification for Internal Tables
The name of an internal table is the internally-unique ID . This
allows you to distinguish between identically-named internal tables from
different procedures . You can find out the
actual name of the internal table by navigating in the ABAP Editor.
ABAP_MODIFICATION_6 Measuring Particular Statements
In LOOP AT itab , the total runtime of all loop passes is
determined. CREATE OBJECT is only measured if the method
CONSTRUCTOR is overwritten.
ABAP_MODIFICATION_7 Hit Lists
In the hit list of the database tables, the buffering is empty if
buffering is switched off or not allowed. As well as the database hit
list, there are object-based hit lists for classes, instances, and
internal tables. For classes and instances, there is a further division
between methods and events.
ABAP_MODIFICATION_8 Use of New Toolss
The various hit lists are created using the ABAP List Viewer .
All normal ALV functions are available within the hit list, although
full navigation from the list is not possible.
The ABAP List Viewer is also used to display the performance
statistics. They are calculated using method calls, and the load time
and net runtime of the program are added to them.
ABAP_MODIFICATION_9 Changes to the User Interface
The user interface of the runtime analysis tool has been fully
revised. The new features are represented by four main areas, into which
you can branch from the initial screen, namely:
Executing the runtime analysis
Switching on an external session
Setting the measurement restrictions
Evaluating the measurement results
Documentation extract taken from SAP system, � Copyright SAP AG. All rights reserved