Vince, On Fri, Jul 29, 2011 at 9:09 AM, Vince Weaver <vweav...@eecs.utk.edu> wrote: > Hello > > I know now and then on this list the matter of deterministic events comes > up. > How do you define 'deterministic' here? Is it: Always get the same answer on a deterministic program (assuming user level monitoring only)?
> Previously the only known x86 Deterministic event was > INST_RETIRED:STORES > on Core2 machines. > > Albert Noll reported to me recently that the > BR_INST_RETIRED:CONDITIONAL > instruction is determinsitic on SandyBridge. I've verified this result. > > Also I've found that the same event is deterministic on Nehalem, with > some overcount issues. This event counts any instruction which > opcode starts with "0x0f". This covers conditional branches, but > unfortunately also counts many MMX instructions. Otherwise though > it is deterministic. > > Vince > vweav...@eecs.utk.edu > > ------------------------------------------------------------------------------ > Got Input? Slashdot Needs You. > Take our quick survey online. Come on, we don't ask for help often. > Plus, you'll get a chance to win $100 to spend on ThinkGeek. > http://p.sf.net/sfu/slashdot-survey > _______________________________________________ > perfmon2-devel mailing list > perfmon2-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/perfmon2-devel > ------------------------------------------------------------------------------ BlackBerry® DevCon Americas, Oct. 18-20, San Francisco, CA The must-attend event for mobile developers. Connect with experts. Get tools for creating Super Apps. See the latest technologies. Sessions, hands-on labs, demos & much more. Register early & save! http://p.sf.net/sfu/rim-blackberry-1 _______________________________________________ perfmon2-devel mailing list perfmon2-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/perfmon2-devel
