Actually some of the information (e.g. sun.ci.lastMethod etc.) is not available via the Tracer Plugin but via jcmd.
# use jps to find the PID of your jvm process jps -v # Display perf counters via jcmd by PID, e.g. 22599 jcmd 22599 PerfCounter.print Cheers, Thomas 2016-06-22 11:54 GMT+02:00 Thomas Darimont <[email protected]>: > In (j)visualvm there is the "Tracer-JVM Probes" plug-in which allows you > to glance at some internal information of the compiler infrastructure. > Not quite what you are looking for but I think it points into the right > direction :) > > JVMStat Counters (sun.ci) > sun.ci.compilerThread.0.compiles=2118 > sun.ci.compilerThread.0.method="" > sun.ci.compilerThread.0.time=218574 > sun.ci.compilerThread.0.type=1 > sun.ci.compilerThread.1.compiles=2295 > sun.ci.compilerThread.1.method="" > sun.ci.compilerThread.1.time=48838 > sun.ci.compilerThread.1.type=1 > sun.ci.lastFailedMethod="org/apache/catalina/deploy/WebXml merge" > sun.ci.lastFailedType=1 > sun.ci.lastInvalidatedMethod="" > sun.ci.lastInvalidatedType=0 > sun.ci.lastMethod="org/apache/jsp/test_jsp setSessionValue" > sun.ci.lastSize=464 > sun.ci.lastType=1 > sun.ci.nmethodCodeSize=13122880 > sun.ci.nmethodSize=27225752 > sun.ci.osrBytes=196768 > sun.ci.osrCompiles=189 > sun.ci.osrTime=130282719 > sun.ci.standardBytes=1847119 > sun.ci.standardCompiles=4223 > sun.ci.standardTime=1098377067 > sun.ci.threads=2 > sun.ci.totalBailouts=1 > sun.ci.totalCompiles=4412 > sun.ci.totalInvalidates=0 > > other counters like "sun.rt" give you some other interesting information > like safepoints, safepointTime etc. > > Cheers, > Thomas > > 2016-06-22 11:37 GMT+02:00 Richard Warburton <[email protected]> > : > >> Hi, >> >> I think a question that should be asked, in the spirit of constructive >> analysis, is why do you want to understand some low level performance >> metric. For example I wouldn't want to look into metrics around say how >> many of my GC cycles are full collections without an idea that GC is a >> throughput or latency problem. What is the high level metric that indicates >> that you need to tune the JIT compiler? >> >> I'm not saying that you're necessarily wrong to look here but it does >> seem unusual and I can't help but feel that applications usually have >> things that are simpler to understand and optimize than JIT compiler >> threads. >> >> >> On Tue, Jun 21, 2016 at 6:36 AM, Brian Toal <[email protected]> wrote: >> >>> Hi all, a few JIT related questions. >>> >>> 1) How do I monitor the length of the compiler queues (client and >>> server)? I want to understand the backlog (if any) to determine if the >>> application could benefit from more compilation threads. >>> >>> 2) What other compilation metrics should I be paying close attention to? >>> Is there a metric that exposes the proportion of code that is running >>> interpreted vs native so I have an idea of how warm the system is. >>> >>> - Brian >>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "mechanical-sympathy" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> For more options, visit https://groups.google.com/d/optout. >>> >> >> >> >> -- >> regards, >> >> Richard Warburton >> >> http://insightfullogic.com >> @RichardWarburto <http://twitter.com/richardwarburto> >> >> -- >> You received this message because you are subscribed to the Google Groups >> "mechanical-sympathy" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> For more options, visit https://groups.google.com/d/optout. >> > > -- You received this message because you are subscribed to the Google Groups "mechanical-sympathy" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
