Previous Topic: 6.4.4.2 I/O Intensity Effects

Next Topic: 6.4.4.4 Explicit Queueing Effects

6.4.4.3 DASD Storage Effects


Variation in the placement of data on DASD devices may affect
the measured CPU time in a variety of ways.

If the number of extents in a target DASD file is not exactly
the same each time, additional end-of-extent processing is
required (i.e., when SPACE=(...,RLSE) is specified on a JCL
DD statement).

For VSAM files, internal fragmentation may vary the number of
CI and CA splits that occur during file extension.  CA splits
are especially expensive, both in direct CPU time and in
overhead to carry out the additional I/O.

Processing time for macros such as LINK, LOAD, XCTL, ATTACH,
and BLDL is affected by where the requested module is
located.  For example, CPU time may be less if the module is
in the LPA, and joblibs and steplibs are not used.  Also, if
the Linklist Lookaside (LLA) software is in use, changes in
the list of supported libraries will affect CPU time.

If the measured activity involves the retrieval of many PDS
members, differences in the PDS directory are reflected in
different amounts of processing to search it.

If the measured activity involves creation or update of a
member of a PDS, differences in the PDS directory may also
vary STOW processing time because of the additional I/O
operations that may be necessary to expand or contract the
directory while keeping it compact and in order.