HP Caliper User's Guide

1. The measurement type of the second database listed in the command is the only
one that will be reported in the diff report output. (In other words, the second
database listed is assumed to be the “base” database.)
2. All matching processes (of the measurement type of the second database specifier)
in each database specifier are merged together, with data in the first database
specifier essentially subtracted from data in the second database specifier.
3. The merged process data is diff'ed and reported.
You can create a diff report from two system-wide collection runs.
HP Caliper supports diff reports for all measurements except the ones below:
• cgprof (HP-UX only)
• cpu (HP-UX only)
• cstack
• pmu_trace
• scgprof
Example of How to Use the caliper diff Command
Assume these two measurement runs:
$ caliper fprof -d fp1 cc himom.c
$ caliper fprof -d fp2 cc -c himom.c
And assume that the first collection (fp1) has data for these processes (on Linux, for
example):
gcc (four instances)
collect2 (two instances)
as
cc1
ld
The second collection (fp2) has data for these processes:
gcc (three instances)
as
cc1
To produce a diff report, use this command:
$ caliper diff fp2 fp1
The report that is produced contains this data:
• Differences between the four gcc processes in the first run and the three gcc
processes in the second run
• Differences between the as processes in the two runs
• Differences between the cc1 processes in the two runs
The collect2 and ld process data is ignored, because there is no matching process
in the other database.
Creating Reports from Multiple Databases 155