2. PERFORMANCE REPORTING ANALYSIS › 2.6 PR/SM LPAR Performance Analysis › 2.6.2 PR/SM LPAR Concepts › 2.6.2.2 Controlling the PR/SM LPAR Environment

2.6.2.2 Controlling the PR/SM LPAR Environment

As was outlined in the previous section, PR/SM shares
processor resources among LPARs via weighted prioritization
and by time-slicing.  Both of these controls can, however, be
overridden by the console operator or system programmer.  The
weighting for each LPAR can be dynamically modified, and the
controls over time-slicing can be altered by means of
enabling the wait assist feature and setting a fixed time
slice.  When this is done, PR/SM will not take control of a
CP when the logical processor goes into a wait state, but
leaves the CP under the control of the LPAR for the entire
duration of the time slice.

This level of control exacts a price:  the PR/SM feature
itself uses some of the processor resources that it is
managing and allocating on behalf of the logical partitions.
This overhead is usually quite small and is directly related
to the ratio of active logical processors to the number of
physical processors under PR/SM control.  The use of
dedicated processors greatly reduces this overhead, perhaps
to as low as 0.1 percent.  The overhead can grow to as much
as 12 percent in some environments as more processors are
shared and as contention for their use increases.

LPARs with dedicated processors have such low overhead that
in some circumstances it is possible to increase rather than
decrease the throughput of a complex by using PR/SM.  Since
the overhead involved in managing, coordinating, and
synchronizing multiple processors by a single SCP can be
substantial, reducing or eliminating multiple CPs can result
in substantial savings of processor resources.  If several
workloads currently supported simultaneously within a single
multiple-processor system can be divided into multiple
single-processor LPARs with dedicated CPs, up to a 10 percent
increase in overall throughput is possible.