HP Caliper User Guide Release 5.5 (5900-2351, August 2012)

The queues event set provides bus request queue (BRQ) information that might give insight into
possible performance problems related to the system bus. The BRQ is a centralized queueing
structure that collects almost all requests from the L1 cache and then schedules those requests to
the L2 cache or front side bus (FSB). High values on the available metrics will likely indicate levels
of bus utilization. This can be confirmed with the sysbus event set.
If you use this event set, the default is to make the measurements irrespective of CPU operating
state (that is, user, system, or interrupt states). By default, the idle state is not included in the
measurement. You can use command-line options to limit the scope of the measurement. Specifically,
you can:
• Limit measurement to a specific privilege level: -m event_set[:all|user|kernel]
• Include idle: --exclude-idle False
• Exclude the interruption state: --measure-on-interrupts off
• Only measure the interruption state: --measure-on-interrupts only
Metrics Available from this Measurement
The following metrics are available from this event set. These descriptions do not take into account
any command-line options you might use.
The metrics are:
• CPU Cycles
This is the number of CPU cycles that were observed during the sample period. It is used with
BRQ read requests issued to compute the BRQ requests insert rate. It can also be useful to
determine the percentage of time it is being monitored.
• BRQ Read Requests Inserted
• This is a count of the number of requests that were inserted into the BRQ during the sample
observation period.
• BRQ Read Requests Per Sec
This is the number of BRQ requests inserted per second. It gives a view of the total demand
that a processor is delivering to the system.
• AVG BRQ Live Entries Per Cycle (not present in dual-core Itanium 2 and Itanium 9300 quad-core
processor systems)
This is the average number live BRQ entries on a per-cycle basis. You can use it to obtain an
idea of how the system is responding to load. Values less than 1.0 indicate very light load,
and values approaching the depth of the BRQ queue (16) indicate a system under considerable
stress.
• AVG BRQ Latency
This is the average number of cycles that a request resides in the BRQ. It is also useful for
interpreting system loading. Large values (> 20 cycles) indicate that the processor is being
delayed during bus request arbitration, probably due to excessive bus utilization by a priority
agent (I/O).
• AVG IOQ Live Entries Per Cycle
The in-order queue (IOQ) monitors all outstanding bus transactions generated by any bus
agent. Requests are loaded into the IOQ during the bus transaction request phase. Transactions
are retired from the IOQ on receipt of a positive response status from the bus. This metric
provides an estimate of the number of bus transactions that are live in the IOQ on a per-cycle
basis. A high value for this metric would be indicative of high system bus utilization (many
transactions in flight).
• AVG OOQ Live Entries Per Cycle
If a deferred response is received while a request is live in the IOQ, it is moved to the
out-of-order queue (OOQ). When the response for the deferred transaction is received, the
OOQ entry is moved to the end of the IOQ to complete the processing of the transaction. This
queues Event Set 255