hwpmc & sampling

[Joseph Koshy]

ns> Seems wierd huh? I'd love to get this to work and would'nt
ns> mind investigating the issue if you point me in the right
ns> direction.

Wierd.

The output of 'time' that you posted showed low usr+sys
numbers, but a large elapsed time.  This would happen if
the process is sleeping waiting for data or an external
event.

ns> real    0m24.728s
ns> user    0m0.001s
ns> sys     0m0.004s

If this was a typo and the process was actually using the
CPU for all 24.7 seconds, could you compare the gprof
profiles you get from a pmcstat -S and a pmcstat -P
run?

--
FreeBSD Volunteer,     http://people.freebsd.org/~jkoshy