Figure 7-11. Shown is an example of severe library latch contention when mutexes are not used. While not shown, Oracle is consuming all the available host CPU, and because the load is based on a PL/SQL loop, recursive SQL is consuming over half of the Oracle CPU.

Figure 7-12 is based on the same load and interval as Figure 7-11. The only difference is library cache mutexes have been enabled by setting the instance parameter _kks_use_mutex_pin to true (which is actually the default). Notice the top wait event is cursor: pin S. This is the result of the cursor being repeatedly and intensely opened and closed.

While the percentage of recursive SQL was the same with mutexes enabled and disabled, when latches were used, the total CPU consumption was nearly twice as much! In fact, look closely and compare the time waited for library cache-related events in Figures 7-11 and 7-12. Once again, mutexes are shown to be more efficient than latches.