Hi, I don't remember if there was sampling support in the Cell support. That's not because you can count that you can sample.
Could you try running the libpfm examples_v2.x examples and in particular notify_self? This is a good indicator of sampling support. On Thu, Sep 2, 2010 at 4:03 PM, Daniele G.Spampinato <daniele.spampin...@inf.ethz.ch> wrote: > Dear all, > > I have problems sampling with pfmon 3.9 on Cell (QS22). > > Whenever i try to sample (regardless of the specific event) I get an > empty set of samples: > > ./pfmon --long-smpl-periods=5000 --short-smpl-period=20 --verbose > ../tests/forktest > > measuring at user privilege level ONLY > 1 event set(s) defined > long sampling periods(val/mask/seed): 5000/0x0/0 > short sampling periods(val/mask/seed): 20/0x0/0 > using inst-hist sampling module > unavailable_pmcs=0xff000000 > unavailable_pmds=0xffffff00 > pmd setup for event set0: > [pmd0 set=0 smpl_pmds=0x0 reset_pmds=0x0] > [pmd0 set=0 ival=0xffffffffffffec78 long_rate=0xffffffffffffec78 > short_rate=0xffffffffffffffec mask=0x0 seed=0 randomize=n] > max=4294967295 cur=4294967295 > locked_mem_avail=3973906432 pfm_mem_avail=3973906432 mem_avail=3973906240 > ncpus=0 max_entries=99347656 smpl_entries=2048 orig_smpl_entries=0 pgsz=65536 > hdrsz=128 entrysz=40 slack=64 > sampling buffer #entries=2048 size=82112, max_entry_size=40 > vCPU0 -> pCPU0 > vCPU1 -> pCPU1 > vCPU2 -> pCPU2 > vCPU3 -> pCPU3 > exec-pattern=* > [3492] started task: ../tests/forktest > follow_exec=n follow_vfork=n follow_fork=n follow_pthread=n > [3492] 32-bit binary > [3492] monitoring /export/home/danieles/tmp/pfmon-3.9/tests/forktest... > [3492] activating monitoring > measurements started at Thu Sep 2 15:14:47 2010 > > creating 0 additional process(es) > 1000 iterations > [3492] task exited > [3492] detached > sdesc chain to process: > [3492] (/export/home/danieles/tmp/pfmon-3.9/tests/forktest) > # results for [3492:3492<-[3487]] > (/export/home/danieles/tmp/pfmon-3.9/tests/forktest) > # total samples : 0 > # total buffer overflows : 0 > # > # event00 > # counts %self %cum code addr > > However, it seems like I should be able to use this feature. > If a run > > pfmon -I > > then I have all the sampling modules available: > > detected host CPUs: Cell Broadband Engine. > detected pfmon PMU model: Cell/B.E PMU > detected libpfm PMU model: CELL > max counters/set: 8 > supported pfmon PMU models: [Cell/B.E PMU] > supported sampling modules: [inst-hist] [detailed] [compact] [raw] > pfmlib version: 3.9 > kernel perfmon version: 2.8 > kernel clock resolution: 1ns (1000000000Hz) > host kernel architecture: ppc64 > > Any suggestion that could help me with this problem? > > Thanks and best regards, > > Daniele > > > ------------------------------------------------------------------------------ > This SF.net Dev2Dev email is sponsored by: > > Show off your parallel programming skills. > Enter the Intel(R) Threading Challenge 2010. > http://p.sf.net/sfu/intel-thread-sfd > _______________________________________________ > perfmon2-devel mailing list > perfmon2-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/perfmon2-devel > ------------------------------------------------------------------------------ This SF.net Dev2Dev email is sponsored by: Show off your parallel programming skills. Enter the Intel(R) Threading Challenge 2010. http://p.sf.net/sfu/intel-thread-sfd _______________________________________________ perfmon2-devel mailing list perfmon2-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/perfmon2-devel
