Hi, The perfmon2 projects has been stopped 2 years ago now.
I you are starting a new performance monitoring project, I strongly suggest you use perf_event instead. This new kernel API provides most of the features present in perfmon2 with the advantage of being included in the upstream Linux kernel since 2.6.31. If you use a recent Linux distro, then you have it. The perf_event tool is called perf. It is part of the kernel source tree. But it is also available as a package on some distros. On Wed, Feb 9, 2011 at 9:05 AM, 陳韋任 <che...@iis.sinica.edu.tw> wrote: > Hi, folks > > I have been reading the document "Perfmon2: a flexible performance monitoring > interface for Linux" > (http://www.kernel.org/doc/ols/2006/ols2006v1-pages-269-288.pdf) > , and I have a couple of questions. > > 1. What the document said is still true for all perfmon released since then? > And I also refer to "The perfmon2 interface specification" > (http://www.hpl.hp.com/techreports/ > 2004/HPL-2004-200R1.html). Anything in those documents is changed? > > 2. In section "4.2 System-wide monitoring", > > "The CPU on which the call to pfm_load_context is executed determines the > monitored CPU. ... The affinity can later be modified." > > Does it mean that if we want to monitor CPU_1, we should pin ourself on > CPU_1 then > call pfm_load_context. After that we can move to another CPU and still > monitor CPU_1? > > 3. In section "5 Sampling Support", > > "The perfmon2 interface has an extensive set of features to support sampling. > ... But there is also kernel-level support to minimize the overhead." > > /* > * set the privilege mode: > * PFM_PLM3 : user level > * PFM_PLM0 : kernel level > */ > inp.pfp_dfl_plm = PFM_PLM3; > > I found above code snipt in the example. What it does is the kernel-level > support > sampling? What are differences between user-level and kernel-level sampling? > > 4. In section "5.4 Custom Sampling Buffer Formats", it said that a format > "n-way buffering" > has been released. But I only found one format belows in perfmon_dfl_smpl.h. > > #define PFM_DFL_SMPL_NAME "default > > Where else I can find other formats, or should I upgrade to newer libpfm? > > > Thanks! > > Regards, > chenwj > > -- > Wei-Ren Chen (陳韋任) > Parallel Processing Lab, Institute of Information Science, > Academia Sinica, Taiwan (R.O.C.) > Tel:886-2-2788-3799 #1667 > > ------------------------------------------------------------------------------ > The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE: > Pinpoint memory and threading errors before they happen. > Find and fix more than 250 security defects in the development cycle. > Locate bottlenecks in serial and parallel code that limit performance. > http://p.sf.net/sfu/intel-dev2devfeb > _______________________________________________ > perfmon2-devel mailing list > perfmon2-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/perfmon2-devel > ------------------------------------------------------------------------------ The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE: Pinpoint memory and threading errors before they happen. Find and fix more than 250 security defects in the development cycle. Locate bottlenecks in serial and parallel code that limit performance. http://p.sf.net/sfu/intel-dev2devfeb _______________________________________________ perfmon2-devel mailing list perfmon2-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/perfmon2-devel
