HP Caliper 5.3 for HP-UX Release Notes (5900-1559, February 2011)

6 Changes in Previous Releases
This section describes changes for the three releases prior to this one.
Changes in Release 5.2
New Options and Enhancements
The HP Caliper 5.2 release has the following improvements:
• Support is added for the quad-core Intel® Itanium® processor 9300 series.
• Support is added for simultaneous fprof sampling on multiple PMU counters. The -s or
--sampling-spec option can be used multiple times to specify the list of sampling events.
• New details are added to the Blocking Primitive section of the Call Stack Profile report with
thread synchronization primitive's holder and waiter thread analysis (HP-UX 11.31 only).
• A new option, --percent-sample=threads|process, is added. This option displays
thread details with the percentage data for samples collected for each thread or for a process.
• A new cstack enhancement option, --stop-functions is added to specify the list of
kernel stop functions. This list is used to avoid unnecessary kernel unwinding.
Deprecations
The following Caliper options are deprecated as of this release and will be deleted in a future
release:
• All instrumentation measurements(cgprof, fcover, and fcount)
• --module-default/include/exclude
• --inlines
• --latency-buckets
• --per-module-data
Deletions
None.
Changes in Release 5.1
New Options and Enhancements to Options
The HP Caliper 5.1 release has the following improvements:
• The cstack report has been enhanced to report inter-process blocking primitives (semaphores,
message queues, sockets, pipes, regular file descriptors, etc.). This process level blocking
primitive information is provided in addition to the previously reported primitives (mutexes,
rwlocks, condition variables, etc.).
• A new HP-UX only option --kernel-stack True|False specifies whether to extend the
callstack samples collection into kernel space. By default, only the userspace callstack samples
will be collected. This option is used only with the cstack measurement.
• A new overview measurement provides flat profile, data cache, and call stack profile
information in a single run.
• The collection speed of scgprof (sampled call graph) has been improved.
• Improvements have been made to the storage of performance data during collection.
Changes in Release 5.2 13