On Thu, Jan 7, 2010 at 7:57 AM, Chad Mynhier <cmynh...@gmail.com> wrote: > On Thu, Jan 7, 2010 at 8:44 AM, Paul <paul.mad...@swx.com> wrote: >> I have no problem finding, from with a "C" app, my process utilization. >> >> I need now to dig deeper and extract internally to the app the CPU >> utilisztion of each thread within my app >> >> Any help, better still source fragment, would be much appreciated > > prstat(1M) gives you per-thread statistics with the -L option. That > would probably be a good place to start: > > http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/cmd/prstat/
Note that prstat -L will give you CPU performance data that is time decayed and the % is over all of the CPU's in the system. That is, if you are on a T5240 with 2 x 8 core CPU's, this will be seen as 128 CPU's to solaris. A thread that spends 100% of its time on CPU will show up as using < 1% CPU. When I am looking at process CPU utilization statistics, I use prstat -mL. The following output from prstat -mLcn 3 gives microstate CPU utilization statistics for the 3 busiest threads. Total: 47 processes, 196 lwps, load averages: 0.15, 0.14, 0.13 PID USERNAME USR SYS TRP TFL DFL LCK SLP LAT VCX ICX SCL SIG PROCESS/LWPID 1221 root 2.0 0.6 0.0 0.0 0.0 0.0 97 0.0 98 3 2K 0 pkg/1 1288 root 0.0 0.2 0.0 0.0 0.0 0.0 100 0.0 4 1 163 0 prstat/1 859 jack 0.1 0.1 0.0 0.0 0.0 0.0 100 0.0 23 1 304 0 sshd/1 Total: 47 processes, 196 lwps, load averages: 0.15, 0.14, 0.13 PID USERNAME USR SYS TRP TFL DFL LCK SLP LAT VCX ICX SCL SIG PROCESS/LWPID 1221 root 1.1 0.4 0.0 0.0 0.0 0.0 99 0.0 65 4 1K 0 pkg/1 1288 root 0.0 0.2 0.0 0.0 0.0 0.0 100 0.0 4 0 163 0 prstat/1 859 jack 0.0 0.0 0.0 0.0 0.0 0.0 100 0.0 17 1 192 0 sshd/1 Total: 47 processes, 196 lwps, load averages: 0.14, 0.13, 0.13 With the -m output, 100% means 100% of a single CPU (or core or strand as appropriate). -- Mike Gerdts http://mgerdts.blogspot.com/ _______________________________________________ perf-discuss mailing list perf-discuss@opensolaris.org
