A Collection of JVM Options

A Collection of JVM Options

This document is a compilation of all the JVM options for various versions of the JVM on primarily SPARC/Solaris Platform. The descriptions for each option are taken mostly verbatim from the reference documents.

The JavaTM HotSpot VM Options page makes some good points regarding the different types of options, which should be followed when reading this document as well. To paraphrase:

  • Standard options recognized by the Java HotSpot VM are described on the Java Application Launcher reference pages.
  • Options that begin with -X are non-standard (not guaranteed to be supported on all VM implementations), and are subject to change without notice in subsequent releases of the Java 2 SDK.
  • Because the -XX options have specific system requirements for correct operation and may require privileged access to system configuration parameters, they are not recommended for casual use. These options are also subject to change without notice.

Please submit any comments, corrections, additions to jvm-options-list@sun.com .

Caveats: I am not associated with any Java Platform development group, I simply put this list together from the various reference sources I could find and without referring to any source code. I have done my best to ensure the options are valid for representative JVMs, however I have not tested every version of every JVM so it is quite possible that options may not be valid in all cases.

Option Type 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Description
-client bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Use Client HotSpot VM. Must be first option.
-server bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Use Server HotSpot VM. Must be first option.
-verbose[:class|gc|jni] bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 enable output of various pieces of the VM. specifying options twice may produce more detailed output.
-cp path 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 set search path for application classes and resources
-classpath path 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 set search path for application classes and resources
-version bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 print product version and exit
-showversion bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 print product version and continue
-fullversion bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 print full java version and exit
-agentlib:<libname>[=<options>] nv 1.6.0 load native agent library <libname>, e.g. -agentlib:hprof see also, -agentlib:jdwp=help and -agentlib:hprof=help
-agentpath:<pathname>[=<options>] nv 1.6.0 load native agent library by full pathname
-javaagent:<jarpath>[=<options>] nv 1.6.0 load Java programming language agent, see java.lang.instrument
-d32 bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 use a 32-bit data model if available
-d64 bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 use a 64-bit data model if available
-ea[:packagename...|:classname] list 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 enable assertions
-enableassertions[:packagename...|:classname] list 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 enable assertions
-da[:packagename...|:classname] list 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 disable assertions
-disableassertions[:packagename...|:classname] list 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 disable assertions
-esa bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 enable system assertions
-enablesystemassertions bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 enable system assertions
-dsa bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 disable system assertions
-disablesystemassertions bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 disable system assertions
-Dname=value nv 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 set a system property
-jre-restrict-search bool 1.5.0 1.6.0 include user private JREs in the version search
-jre-no-restrict-search bool 1.5.0 1.6.0 exclude user private JREs in the version search
-Xt bool O O O O O O turn on instruction tracing
-Xtm bool O O O O O O turn on method tracing
-Xbootclasspath[/a|/p]:<path> list 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 set, append to, or prepend to boot class path
-Xdebug bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 enable remote debugging
-Xfuture bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 enable strictest checks, anticipating future default
-Xcheck:jni bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 perform additional checks for JNI functions
-Xshare:off bool 1.6.0 do not attempt to use shared class data
-Xshare:auto bool 1.6.0 use shared class data if possible (default)
-Xshare:on bool 1.6.0 require using shared class data, otherwise fail.
-Xnoclassgc bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 disable class garbage collection
-Xss<size> size 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 set maximum native stack size for any thread
-Xoss<size> size 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 set maximum Java stack size for any thread. Note: This option is useless with HotSpot as HotSpot doesn’t have separate native and Java stacks.
-Xms<size> size 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 set initial Java heap size
-Xmx<size> size 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 set maximum Java heap size
-Xmn<size> size 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Size of young generation
-Xrs bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 reduce the use of OS signals
-Xrunhprof[:help]|[:option=value, ...] list 1.3.1 1.4.0 1.4.2 1.5.0 1.6.0 perform heap or cpu profiling
-Xmaxjitcodesize<size> size 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 set the maximum size (in bytes) for the JIT code area
-Xsqnopause bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 do not pause for user interaction on SIGQUIT
-Xoptimize bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Use optimizing JIT compiler (SPARC only). Replaced by -server in later VMs
-Xmixed bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 mixed mode execution (default)
-Xint bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 interpreted mode execution only
-Xincgc bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Incremental GC
-Xconcgc bool 1.4.1 1.4.2 1.5.0 1.6.0 Concurrent GC. Replaced by -XX:+UseConcMarkSweepGC (available beginning with J2SE 1.4.1)
-Xprof bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Precursor to -Xrunprof.
-Xaprof path 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Simple allocation profiler. unstable. debug only.
-Xloggc:<file> path 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Log garbage collection actions to a specified file.
-Xcomp bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Forces HotSpot to compile 100% of the code with maximum optimization.
-Xbatch bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Forces HotSpot to compile methods in the foreground. Normally method compilation is moved from the foreground to a background process if they take long. This allows allows the interpreted version to continue to run until the compiled version is complete. This option prevents background compiling.
-Xconcurrentio bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 The main feature turned on with -Xconcurrentio is to use LWP based synchronization instead of thread based synchronization. We have found certain applications to speed up by over 40%. In 1.4, LWP based synchronization is the default, but -Xconcurrentio can still help since it turns on some other internal options.
-Xgenconfig bool 1.6.0 1.2 option only
-Xverifyheap bool O O O O O Verify heap integrity. Replaced by -XX:+VerifyBeforeGC -XX:+VerifyAfterGC -XX:+VerifyTLE (-XX:+VerifyTLAB in J2SE 1.4) -XX:+VerifyBeforeScavenge -XX:+VerifyAfterScavenge (all debug only)
-XX:AdaptivePermSizeWeight=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:AdaptiveSizeDecrementScaleFactor=<value> uintx 1.5.0 1.6.0
-XX:AdaptiveSizePolicyCollectionCostMargin=<value> uintx 1.5.0 1.6.0
-XX:AdaptiveSizePolicyFootprintThreshold=<value> uintx 1.5.0
-XX:AdaptiveSizePolicyInitializingSteps=<value> uintx 1.5.0 1.6.0
-XX:AdaptiveSizePolicyWeight=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:AdaptiveSizeThroughPutPolicy=<value> uintx 1.5.0 1.6.0
-XX:AdaptiveTimeWeight=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+AdjustConcurrency bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+AggressiveHeap bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 This option instructs the JVM to push memory use to the limit: the overall heap is more than 3850MB, the allocation area of each thread is 256K, the memory management policy defers collection as long as possible, and (beginning with J2SE 1.3.1_02) some GC activity is done in parallel.
-XX:+AggressiveOpts bool 1.6.0
-XX:AliasLevel=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:AllocatePrefetchDistance=<value> intx 1.6.0
-XX:AllocatePrefetchInstr=<value> intx 1.6.0
-XX:AllocatePrefetchLines=<value> intx 1.6.0
-XX:AllocatePrefetchStepSize=<value> intx 1.6.0
-XX:AllocatePrefetchStyle=<value> intx 1.6.0
-XX:+AllowJNIEnvProxy bool 1.6.0
-XX:+AllowUserSignalHandlers bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Do not complain if the application installs signal handlers (Solaris Only)
-XX:AltStackSize=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 Alternate signal stack size (in Kbytes, Solaris Only)
-XX:+AlwaysActAsServerClassMachine bool 1.6.0
-XX:+AlwaysCompileLoopMethods bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:AlwaysInflate=<value> intx 1.6.0
-XX:+AlwaysPreTouch bool 1.6.0
-XX:+AlwaysTenure bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:AppendRatio=<value> intx 1.5.0 1.6.0
-XX:+BackgroundCompilation bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Enables or disables JIT compilation in the background. When enabled, interpreted methods are still executed until background compilation is complete.
-XX:BaseFootPrintEstimate=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:BCEATraceLevel=<value> intx 1.6.0
-XX:BiasedLockingBulkRebiasThreshold=<value> intx 1.6.0
-XX:BiasedLockingBulkRevokeThreshold=<value> intx 1.6.0
-XX:BiasedLockingDecayTime=<value> intx 1.6.0
-XX:BiasedLockingStartupDelay=<value> intx 1.6.0
-XX:+BindCMSThreadToCPU bool 1.6.0
-XX:+BindGCTaskThreadsToCPUs bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+BindScavengeThreads bool 1.3.1
-XX:+BlockOffsetArrayUseUnallocatedBlock bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+BytecodeVerificationLocal bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+BytecodeVerificationRemote bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+CheckJNICalls bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+CheckUnsafeOps bool 1.4.0 1.4.1 1.4.2 1.5.0
-XX:CICompilerCount=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+CICompilerCountPerCPU bool 1.4.2 1.5.0 1.6.0
-XX:+CIDynamicCompilePriority bool 1.3.1 1.4.0 1.4.1
-XX:+CIMaxPriority bool 1.4.0 1.4.1
-XX:+CITime bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Time spent in JIT Compiler
-XX:+ClassUnloading bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:ClearFPUAtPark=<value> intx 1.6.0
-XX:+ClipInlining bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:CMSAbortablePrecleanMinWorkPerIteration=<value> uintx 1.5.0 1.6.0
-XX:CMSAbortablePrecleanWaitMillis=<value> intx 1.5.0 1.6.0
-XX:+CMSAbortSemantics bool 1.6.0
-XX:CMSBitMapYieldQuantum=<value> uintx 1.4.2 1.5.0
-XX:CMSBootstrapOccupancy=<value> intx 1.5.0 1.6.0
-XX:+CMSClassUnloadingEnabled bool 1.4.1 1.4.2 1.5.0 1.6.0 It has been reported that for 1.5, this apparently needs to be combined with -XX:+CMSPermGenSweepingEnabled in order to take effect.
-XX:+CMSCompactWhenClearAllSoftRefs bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+CMSConcurrentMTEnabled bool 1.6.0
-XX:CMSDictionaryChoice=<value> intx 1.4.0
-XX:CMSExpAvgFactor=<value> uintx 1.4.2 1.5.0 1.6.0
-XX:CMSFullGCsBeforeCompaction=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0 Parameter that affects compaction of the old generation. If at least this number of concurrent collections has not succeeded between full collection, do a compaction on full collections. If 0, always do compactions on full collections when UseCMSCompactAtFullCollection is true
-XX:CMSIncrementalDutyCycle=<value> uintx 1.4.2 1.5.0 1.6.0
-XX:CMSIncrementalDutyCycleMin=<value> uintx 1.4.2 1.5.0 1.6.0
-XX:+CMSIncrementalMode bool 1.4.2 1.5.0 1.6.0
-XX:CMSIncrementalOffset=<value> uintx 1.4.2 1.5.0 1.6.0
-XX:+CMSIncrementalPacing bool 1.4.2 1.5.0 1.6.0
-XX:CMSIncrementalSafetyFactor=<value> uintx 1.4.2 1.5.0 1.6.0
-XX:CMSIndexedFreeListReplenish=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:CMSInitiatingOccupancyFraction=<value> intx 1.4.1 1.4.2 1.5.0 1.6.0 Sets the threshold percentage of the used heap in the old generation at which the CMS collection takes place. For example, if set to 60, the CMS collector will be initiated every time the old generation becomes 60% full. By default, this threshold is calculated at run time, and the CMS collector might be triggered only when the old generation heap is about 80-90% full. Tuning this value can improve performance in many cases. Since the CMS collector does not suspend the mutators as it sweeps and frees memory, setting this switch can ensure that enough free memory is available for object promotion from the young generation as more data is allocated by the application. Sometimes, if this switch is not tuned, CMS collection might not be able to keep up and may fail, eventually triggering the default stop-the-world mark-compact collector. Available from J2SE1.4.1.
-XX:+CMSLoopWarn bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:CMSMarkStackSize=<value> uintx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:CMSMarkStackSizeMax=<value> uintx 1.5.0 1.6.0
-XX:CMSMaxAbortablePrecleanLoops=<value> uintx 1.5.0 1.6.0
-XX:CMSMaxAbortablePrecleanTime=<value> intx 1.5.0 1.6.0
-XX:+CMSNoWarning bool 1.4.0 1.4.1
-XX:+CMSParallelRemarkEnabled bool 1.4.2 1.5.0 1.6.0 Attempt to descrease remark pauses when used with -XX:+UseParNewGC.
-XX:+CMSParallelSurvivorRemarkEnabled bool 1.6.0
-XX:+CMSPermGenPrecleaningEnabled bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+CMSPermGenSweepingEnabled bool 1.4.1 1.4.2 1.5.0 1.6.0 It has been reported that for 1.5, this apparently needs to be combined with -XX:+CMSClassUnloadingEnabled in order to take effect.
-XX:+CMSPLABRecordAlways bool 1.6.0
-XX:CMSPrecleanDenominator=<value> uintx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+CMSPrecleaningEnabled bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:CMSPrecleanIter=<value> uintx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:CMSPrecleanNumerator=<value> uintx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+CMSPrecleanRefLists1 bool 1.5.0 1.6.0
-XX:+CMSPrecleanRefLists2 bool 1.5.0 1.6.0
-XX:+CMSPrecleanSurvivors1 bool 1.6.0
-XX:+CMSPrecleanSurvivors2 bool 1.6.0
-XX:CMSPrecleanThreshold=<value> uintx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:CMSRescanMultiple=<value> uintx 1.4.2 1.5.0 1.6.0
-XX:CMSRevisitStackSize=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:CMSSamplingGrain=<value> uintx 1.5.0 1.6.0
-XX:+CMSScavengeBeforeRemark bool 1.6.0
-XX:CMSScheduleRemarkEdenPenetration=<value> uintx 1.5.0 1.6.0
-XX:CMSScheduleRemarkEdenSizeThreshold=<value> uintx 1.5.0 1.6.0
-XX:CMSScheduleRemarkSamplingRatio=<value> uintx 1.5.0 1.6.0
-XX:CMSTriggerRatio=<value> intx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+CMSUseOldDefaults bool 1.6.0
-XX:CMSWaitDuration=<value> intx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:CMSWorkQueueDrainThreshold=<value> uintx 1.4.2 1.5.0 1.6.0
-XX:+CMSYield bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:CMSYieldQuantum=<value> intx 1.4.0 1.4.1
-XX:CMSYoungGenPerWorker=<value> intx 1.6.0
-XX:CodeCacheExpansionSize=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:CodeCacheMinimumFreeSpace=<value> uintx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+CodeForP6 bool 1.3.1
-XX:+CollectGen0First bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+CompactFields bool 1.6.0
-XX:CompilationPolicyChoice=<value> intx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:CompilationRepeat=<value> intx 1.4.2 1.5.0
-XX:CompileCommand=<value> ccstr 1.4.2 1.5.0 1.6.0
-XX:CompileCommandFile=<value> ccstr 1.4.2 1.5.0 1.6.0 Allows you to specify alternate compiler command files. When used, this overrides the default behavior of the VM and the VM will NOT scan either the libjvm.sl directory or the current directory for a .hotspot_compiler file.
-XX:CompileOnly=<value> ccstr 1.4.2 1.5.0 1.6.0
-XX:+CompilerThreadHintNoPreempt bool 1.4.2 1.5.0 1.6.0
-XX:CompilerThreadPriority=<value> intx 1.4.2 1.5.0 1.6.0
-XX:CompilerThreadStackSize=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:CompileThreshold=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 number of method invocations/branches before (re-)compiling [10,000 -server, 1,500 -client]
-XX:+ConvertSleepToYield bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+ConvertYieldToSleep bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+DebugNonSafepoints bool 1.6.0
-XX:DefaultInitialRAMFraction=<value> uintx 1.5.0 1.6.0
-XX:DefaultMaxRAM=<value> uintx 1.5.0 1.6.0
-XX:DefaultMaxRAMFraction=<value> uintx 1.5.0 1.6.0
-XX:DefaultThreadPriority=<value> intx 1.4.2 1.5.0 1.6.0
-XX:DefaultTickInterval=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 O
-XX:+DefaultUseSpecLock bool 1.3.1 1.4.0
-XX:DeferPollingPageLoopCount=<value> intx 1.6.0
-XX:DeferThrSuspendLoopCount=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:DelayTickAdjustment=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 O
-XX:+DisableAttachMechanism bool 1.6.0
-XX:+DisableExplicitGC bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Disable calls to System.gc(), JVM still performs garbage collection when necessary
-XX:+DisplayVMOutput bool 1.4.2 1.5.0 1.6.0
-XX:+DisplayVMOutputToStderr bool 1.5.0 1.6.0
-XX:+DisplayVMOutputToStdout bool 1.5.0 1.6.0
-XX:+DontCompileHugeMethods bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+DontYieldALot bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+DontYieldALotInterval bool Interval in milliseconds between yields. (debug only)
-XX:+DTraceAllocProbes bool 1.6.0
-XX:+DTraceMethodProbes bool 1.6.0
-XX:+DTraceMonitorProbes bool 1.6.0
-XX:+DumpSharedSpaces bool 1.5.0 1.6.0
-XX:+EagerInitialization bool 1.4.1
-XX:+EagerXrunInit bool 1.4.2 1.5.0 1.6.0
-XX:EmitLFence=<value> intx 1.6.0
-XX:EmitSync=<value> intx 1.6.0
-XX:+EnableJVMPIInstructionStartEvent bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Enables or disables the JVMPI event, JVMPI_EVENT_INSTRUCTION_START, which is sent when each instruction is issued by the interpreter
-XX:ErrorFile=<value> ccstr 1.6.0
-XX:+EstimateArgEscape bool 1.6.0
-XX:EventLogLength=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+ExitOnBadLibThread bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0
-XX:+ExplicitGCInvokesConcurrent bool 1.6.0
-XX:+ExtendedDTraceProbes bool 1.6.0
-XX:+FailOverToOldVerifier bool 1.6.0
-XX:+FasterSync bool 1.3.1
-XX:+FastNMethodDependencies bool 1.4.1 1.4.2 1.5.0
-XX:+FastTLABRefill bool 1.5.0
-XX:FieldsAllocationStyle=<value> intx 1.6.0
-XX:+FilterSpuriousWakeups bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:FloatCacheThreshold=<value> intx 1.4.0
-XX:+FLSVerifyAllHeapReferences bool 1.6.0
-XX:+FLSVerifyIndexTable bool 1.6.0
-XX:+FLSVerifyLists bool 1.6.0
-XX:+ForceSharedSpaces bool 1.5.0 1.6.0
-XX:+ForceTimeHighResolution bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+ForceUseSpecLock bool 1.3.1 1.4.0
-XX:FreqInlineSize=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Integer specifying maximum number of bytecode instructions in a frequently executed method which gets inlined.
-XX:+FullSpeedJVMDI bool 1.4.0 1.4.1 1.4.2
-XX:GCHeapFreeLimit=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0 Used with -XX:+UseParallelGC to help prevent out-of-memory exceptions from occuring during garbage collection. This parameter specifies the lower limit on the amount of space freed during a garbage collection in per cent of the maximum heap (default is 5).
-XX:GCTaskTimeStampEntries=<value> uintx 1.4.2 1.5.0 1.6.0
-XX:GCTimeLimit=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0 Used with -XX:+UseParallelGC to help prevent out-of-memory exceptions from occuring during garbage collection. This parameter specifies the upper limit on the amount of time spent in garbage collection in per cent of total time (default is 90)
-XX:GCTimeRatio=<value> uintx 1.5.0 1.6.0 The ratio of GC time to application time, ratio being: 1 / (1 + nnn). For example -XX:GCTimeRatio=19 sets a goal of 5% of the total time for GC.
-XX:+HandlePromotionFailure bool 1.5.0 1.6.0
-XX:hashCode=<value> intx 1.6.0
-XX:+HeapDumpOnOutOfMemoryError bool 1.6.0
-XX:HeapDumpPath=<value> ccstr 1.6.0
-XX:HPILibPath=<value> ccstr 1.6.0
-XX:InitialCodeCacheSize=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:InitialSurvivorRatio=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:InitialTenuringThreshold=<value> intx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+Inline bool 1.3.1 1.4.1 1.4.2 1.5.0 1.6.0
-XX:InlineSmallCode=<value> intx 1.3.1 1.4.2
-XX:+InlineUnreachedCalls bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0
-XX:+InstallMethods bool 1.3.1
-XX:InterpreterProfilePercentage=<value> intx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+JavaMonitorsInStackTrace bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Useful to display more stack trace information about monitors
-XX:JavaPriority1_To_OSPriority=<value> intx 1.4.2 1.5.0 1.6.0
-XX:JavaPriority10_To_OSPriority=<value> intx 1.4.2 1.5.0 1.6.0
-XX:JavaPriority2_To_OSPriority=<value> intx 1.4.2 1.5.0 1.6.0
-XX:JavaPriority3_To_OSPriority=<value> intx 1.4.2 1.5.0 1.6.0
-XX:JavaPriority4_To_OSPriority=<value> intx 1.4.2 1.5.0 1.6.0
-XX:JavaPriority5_To_OSPriority=<value> intx 1.4.2 1.5.0 1.6.0
-XX:JavaPriority6_To_OSPriority=<value> intx 1.4.2 1.5.0 1.6.0
-XX:JavaPriority7_To_OSPriority=<value> intx 1.4.2 1.5.0 1.6.0
-XX:JavaPriority8_To_OSPriority=<value> intx 1.4.2 1.5.0 1.6.0
-XX:JavaPriority9_To_OSPriority=<value> intx 1.4.2 1.5.0 1.6.0
-XX:+JNIDetachReleasesMonitors bool 1.6.0
-XX:+JVMPICheckGCCompatibility bool 1.5.0 1.6.0
-XX:LargePageHeapSizeThreshold=<value> uintx 1.5.0 1.6.0
-XX:LargePageSizeInBytes=<value> uintx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+LazyBootClassLoader bool 1.6.0
-XX:+LIRCacheLocals bool 1.4.0 1.4.1 1.4.2 1.5.0
-XX:+LIRFillDelaySlots bool 1.4.0 1.4.1 1.4.2 1.5.0
-XX:+LIROptimize bool 1.4.0
-XX:+LIROptimizeDeleteOps bool 1.4.0
-XX:+LIROptimizeFloats bool 1.4.0
-XX:+LIROptimizeStack bool 1.4.0
-XX:LocalCacheThreshold=<value> intx 1.4.0
-XX:+LogCompilation bool 1.4.2 1.5.0 1.6.0
-XX:LogFile=<value> ccstr 1.4.2 1.5.0
-XX:+LogVMOutput bool 1.4.2 1.5.0 1.6.0
-XX:MallocVerifyInterval=<value> intx 1.6.0
-XX:MallocVerifyStart=<value> intx 1.6.0
-XX:+ManagementServer bool 1.5.0 1.6.0
-XX:MarkSweepAlwaysCompactCount=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:MarkSweepDeadRatio=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:MaxBCEAEstimateLevel=<value> intx 1.6.0
-XX:MaxBCEAEstimateSize=<value> intx 1.6.0
-XX:MaxDirectMemorySize=<value> intx 1.4.2 1.5.0 1.6.0
-XX:+MaxFDLimit bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Bump the number of file descriptors to max. (Solaris only)
-XX:MaxGCMinorPauseMillis=<value> uintx 1.5.0 1.6.0
-XX:MaxGCPauseMillis=<value> uintx 1.5.0 1.6.0 A hint to the virtual machine that pause times of nnn milliseconds or less are desired. The vm will adjust the java heap size and other gc-related parameters in an attempt to keep gc-induced pauses shorter than nnn milliseconds. Note that this may cause the vm to reduce overall throughput, and in some cases the vm will not be able to meet the desired pause time goal.
-XX:MaxHeapFreeRatio=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 heap free percentage (default 70)
-XX:MaxHeapSize=<value> uintx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:MaxInlineLevel=<value> intx 1.3.1 1.4.2
-XX:MaxInlineSize=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Integer specifying maximum number of bytecode instructions in a method which gets inlined.
-XX:MaxJavaStackTraceDepth=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:MaxLiveObjectEvacuationRatio=<value> uintx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:MaxNewSize=<value> uintx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Maximum size of new generation (in bytes) [32m sparc, 2.5m intel for 1.3, no limit for 1.4 as NewRatio is now used to determine MaxNewSize]
-XX:MaxPermHeapExpansion=<value> uintx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:MaxPermSize=<value> uintx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Size of the Permanent Generation
-XX:MaxRecursiveInlineLevel=<value> intx 1.3.1 1.4.2
-XX:MaxSurvivorShrinkRatio=<value> size_t 1.3.1
-XX:MaxTenuringThreshold=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 This switch determines how much the objects may age in the young generation before getting promoted to the older generation. The default value is 31. For a big enough young generation and “survivor space”, the long-lived objects may be copied up to 31 times between the survivor spaces before they are finally promoted to the old generation.
-XX:MaxTickInterval=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 O
-XX:MaxTLABRatio=<value> uintx 1.4.0 1.4.1 1.4.2
-XX:MaxTLERatio=<value> 1.3.1 1.6.0
-XX:MinHeapDeltaBytes=<value> uintx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:MinHeapFreeRatio=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 heap free percentage (default 40)
-XX:MinOldHeapExpansion=<value> size_t 1.3.1
-XX:MinPermHeapExpansion=<value> uintx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:MinSurvivorRatio=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:MinTickInterval=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 O
-XX:MinTLABSize=<value> uintx 1.5.0 1.6.0
-XX:NativeSynchronization=<value> ccstr 1.5.0
-XX:+NeverActAsServerClassMachine bool 1.6.0
-XX:+NeverTenure bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:NewRatio=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Ratio of new/old generation sizes [sparc -server: 2, sparc -client: 4 (1.3) 8 (1.3.1+), intel: 12]
-XX:NewSize=<value> uintx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Default size of new generation (in bytes) [sparc 2.125M, intel: 640k]
-XX:NewSizeThreadIncrease=<value> uintx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 As more threads are created in a server application, the object allocation rate may increase with the number of active threads. The number of active threads is considered when adjusting the size of the young space, after a garbage collection. This flag specifies, in Kilobytes, the increment in young object space size, per active thread, to accomodate potentially faster object allocation rate.
-XX:NmethodSweepFraction=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:NMethodSweepInterval=<value> 1.3.1
-XX:NUMAChunkResizeWeight=<value> intx 1.6.0
-XX:NUMAPageScanRate=<value> intx 1.6.0
-XX:NUMASpaceResizeRate=<value> intx 1.6.0
-XX:+NUMAStats bool 1.6.0
-XX:OldPLABSize=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:OldSize=<value> uintx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+OmitStackTraceInFastThrow bool 1.6.0
-XX:OnError=<value> ccstr 1.5.0 1.6.0 The Hotspot JVM includes a fatal error hander that can run a user-supplied script or program if the JVM aborts. A debug tool can also connect to a hung JVM or core file using the Hotspot JVM serviceability agent connector.
-XX:OnOutOfMemoryError=<value> ccstr 1.6.0
-XX:OnStackReplacePercentage=<value> intx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:OnStackReplaceThreshold=<value> 1.3.1
-XX:+OptimizeSinglePrecision bool 1.4.0
-XX:+OverrideDefaultLibthread bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 On Solaris 9, this option is not necessary. On Solaris 8, J2SE? versions 1.3.1_02+ and 1.4+ require this option when using the alternate threads library. This option is not possible on pre-Solaris 8 operating environments.
-XX:OversizedCarThreshold=<value> uintx 1.4.0 1.4.1 1.4.2 1.5.0 O
-XX:ParallelGCBufferWastePct=<value> intx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:ParallelGCG1AllocBufferSize=<value> intx 1.4.1 1.4.2 1.5.0
-XX:ParallelGCOldGenAllocBufferSize=<value> intx 1.4.1 1.4.2 1.5.0 1.6.0 Java 6: se -XX:OldPLABSize in place of -XX:ParallelGCOldGenAllocBufferSize in the future.
-XX:+ParallelGCRetainPLAB bool 1.5.0 1.6.0
-XX:ParallelGCTermSpinLoopMS=<value> intx 1.4.1 1.4.2 1.5.0
-XX:ParallelGCThreads=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0 This switch sets the number of parallel GC threads that the JVM must run for performing garbage collection in the young generation. The default value of n is equal to the number of CPUs on the system. However, it has been observed that tweaking this number may improve performance in some cases. An example might be that of multiple instances of JVM running on a single multi-CPU system. In this case, the number of parallel GC threads for each JVM may need to be set to less than the number of CPUs by using this switch. Available from J2SE1.4.1.
-XX:ParallelGCToSpaceAllocBufferSize=<value> intx 1.4.1 1.4.2 1.5.0 1.6.0 Java 6: Use -XX:YoungPLABSize in place of -XX:ParallelGCToSpaceAllocBufferSize in the future.
-XX:+ParallelGCVerbose bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+ParallelRefProcEnabled bool 1.5.0 1.6.0
-XX:ParCMSPromoteBlocksToClaim=<value> intx 1.4.1 1.4.2 1.5.0 1.6.0 Java 6: Use -XX:CMSParPromoteBlocksToClaim in place of -XX:ParCMSPromoteBlocksToClaim in the future.
-XX:ParGCArrayScanChunk=<value> intx 1.5.0 1.6.0
-XX:ParGCDesiredObjsFromOverflowList=<value> intx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:ParGCMaxArrayAsNormalObject=<value> intx 1.4.1 1.4.2
-XX:+ParkTLE bool 1.3.1
-XX:+PassParametersInRegisters bool 1.3.1
-XX:+PauseAtStartup bool 1.6.0
-XX:PauseAtStartupFile=<value> ccstr 1.6.0
-XX:PerBytecodeRecompilationCutoff=<value> intx 1.6.0
-XX:PerBytecodeTrapLimit=<value> intx 1.6.0
-XX:+PerfAllowAtExitRegistration bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+PerfBypassFileSystemCheck bool 1.4.2 1.5.0 1.6.0
-XX:PerfDataMemorySize=<value> intx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:PerfDataSamplingInterval=<value> intx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+PerfDataSaveToFile bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+PerfDisableSharedMem bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:PerfMaxStringConstLength=<value> intx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:PerMethodRecompilationCutoff=<value> intx 1.6.0
-XX:PerMethodTrapLimit=<value> intx 1.6.0
-XX:PermGenPadding=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:PermMarkSweepDeadRatio=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:PermSize=<value> uintx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Initial size of permanent generation
-XX:+PostSpinYield bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:PreBlockSpin=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+PreferInterpreterNativeStubs bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:PrefetchCopyIntervalInBytes=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Prefetch n bytes ahead for scavenge copy destination area. The default is 0. The ideal value may vary depending on the processor.
-XX:PrefetchFieldsAhead=<value> intx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:PrefetchScanIntervalInBytes=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Prefetch n bytes ahead for scavenge scan area. The default is 0. The ideal value may vary depending on the processor.
-XX:PrefetchScavengeFieldsAhead=<value> 1.3.1 Prefetch the next n fields of objects during scavenge. The default is 0. The ideal value may vary depending on the processor.
-XX:PreInflateSpin=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+PreserveAllAnnotations bool 1.5.0 1.6.0
-XX:PreserveMarkStackSize=<value> uintx 1.5.0 1.6.0
-XX:+PreSpinYield bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:PretenureSizeThreshold=<value> uintx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Objects size greater than this is directly allocated in the older generation
-XX:+PrintAdaptiveSizePolicy bool 1.4.1 1.4.2 1.5.0 1.6.0 Prints information about adaptive size policy
-XX:+PrintClassHistogram bool 1.4.2 1.5.0 1.6.0 Prints the all the java heap objects, their instance count and total space they occupy in the heap. The only downside is that you need to issue a SIGQUIT (see -Xsqnopause) which will leave the app running but will dump all of this data to stdout. Very useful to assist in identifying memory problems for example on a production platform where an CPU intensive profiler cannot be used.
-XX:+PrintCMSInitiationStatistics bool 1.6.0
-XX:PrintCMSStatistics=<value> intx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 If > 0, Print statistics about the concurrent collections. For example, the number of times the concurrent collection yield to a young generation collection and the number of cards precleaned
-XX:+PrintCommandLineFlags bool 1.5.0 1.6.0
-XX:+PrintCompilation bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 traces methods as compiled
-XX:+PrintConcurrentLocks bool 1.6.0
-XX:PrintFLSCensus=<value> intx 1.4.1 1.4.2 1.5.0 1.6.0 if > 0, print the populations of the CMS free lists
-XX:PrintFLSStatistics=<value> intx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 If > 0, print statistics about the concurrent free lists. For example, a fragmentation parameter
-XX:+PrintGC bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+PrintGCApplicationConcurrentTime bool 1.4.2 1.5.0 1.6.0 Measure the amount of time the applications runs between collection pauses
-XX:+PrintGCApplicationStoppedTime bool 1.4.2 1.5.0 1.6.0 Measure the length of the collection pauses.
-XX:+PrintGCDetails bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Gives some details about the GCs, such as size of the young and old generation before and after GCs, size of total heap, time it takes for a GC to happen in young and old generation, size of objects promoted at every GC etc.
-XX:+PrintGCTaskTimeStamps bool 1.4.2 1.5.0 1.6.0
-XX:+PrintGCTimeStamps bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Prints the times at which the GCs happen relative to the start of the application
-XX:+PrintHeapAtGC bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Prints detailed GC info including heap occupancy before and after GC.
-XX:+PrintHeapAtSIGBREAK bool 1.6.0
-XX:+PrintHeapUsageOverTime bool Print heap usage and capacity with timestamps
-XX:+PrintInlining bool Print message about inlined methods (debug only)
-XX:+PrintInterpreter bool 1.6.0
-XX:+PrintJNIResolving bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+PrintParallelOldGCPhaseTimes bool 1.6.0
-XX:+PrintParScavenge bool 1.3.1
-XX:+PrintPLAB bool 1.6.0
-XX:+PrintScavenge bool 1.3.1
-XX:+PrintSharedSpaces bool 1.5.0 1.6.0
-XX:+PrintTenuringDistribution bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Gives the aging distribution of the allocated objects in the young generation. Tuning of -XX:NewSize, -XX:MaxNewSize, -XX:SurvivorRatio and -XX:MaxTenuringThreshold=0as described earlier, should be directed by the analysis of the output from this switch to determine that objects are not prematurely promoted to the old generation.
-XX:+PrintTLAB bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 trace all the operations on TLAB’s (Thread Allocation Buffers)
-XX:+PrintTLE bool 1.3.1 1.6.0
-XX:+PrintVMOptions bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+PrintVMQWaitTime bool 1.6.0
-XX:ProcessDistributionStride=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:ProcessingToTenuringRatio=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 O
-XX:+ProfileIntervals bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:ProfileIntervalsTicks=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:ProfileMaturityPercentage=<value> intx 1.6.0
-XX:+ProfilerLight bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0
-XX:+ProfilerPrintByteCodeStatistics bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+ProfilerRecordPC bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+ProfileVM bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:PromotedPadding=<value> uintx 1.5.0 1.6.0
-XX:QueuedAllocationWarningCount=<value> intx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+RangeCheckElimination bool 1.6.0
-XX:ReadPrefetchInstr=<value> intx 1.6.0
-XX:ReadSpinIterations=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+ReduceSignalUsage bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:RefDiscoveryPolicy=<value> intx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+RegisterFinalizersAtInit bool 1.6.0
-XX:+RelaxAccessControlCheck bool 1.6.0
-XX:+RequireSharedSpaces bool 1.5.0 1.6.0
-XX:ReservedCodeCacheSize=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Reserved code cache size (in bytes) – maximum code cache size. [Solaris 64-bit: 1024m]
-XX:+ResizePLAB bool 1.6.0
-XX:+ResizeTLAB bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Enable or disable resizing of TLABs. The -server VM has this option enabled by default.
-XX:+ResizeTLE bool 1.3.1 1.6.0
-XX:+RestoreMXCSROnJNICalls bool 1.6.0
-XX:+RewriteBytecodes bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+RewriteFrequentPairs bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+SafepointPolling bool 1.5.0
-XX:SafepointSpinBeforeYield=<value> intx 1.6.0
-XX:+ScavengeBeforeFullGC bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:ScavengeThreads=<value> 1.3.1
-XX:SelfDestructTimer=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+SerializeVMOutput bool 1.4.2 1.5.0 1.6.0
-XX:SharedDummyBlockSize=<value> uintx 1.5.0 1.6.0
-XX:SharedMiscDataSize=<value> uintx 1.5.0 1.6.0
-XX:+SharedOptimizeColdStart bool 1.6.0
-XX:SharedReadOnlySize=<value> uintx 1.5.0 1.6.0
-XX:SharedReadWriteSize=<value> uintx 1.5.0 1.6.0
-XX:+SharedSkipVerify bool 1.6.0
-XX:+ShowMessageBoxOnError bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:SliceSizeInCards=<value> 1.3.1
-XX:SoftRefLRUPolicyMSPerMB=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Starting with Java HotSpot VM implementations in J2SE 1.3.1, softly reachable objects will remain alive for some amount of time after the last time they were referenced. The default value is one second of lifetime per free megabyte in the heap. This value can be adjusted using the -XX:SoftRefLRUPolicyMSPerMB flag, which accepts integer values representing milliseconds per MB of free memory.
-XX:+SplitIfBlocks bool 1.6.0
-XX:StackRedPages=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:StackShadowPages=<value> intx 1.4.2 1.5.0 1.6.0
-XX:+StackTraceInThrowable bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Enable or disable stack traces in exceptions.
-XX:StackYellowPages=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:Stage2CompileThreshold=<value> intx 1.4.0
-XX:+StagedCompilation bool 1.4.0
-XX:+StartAttachListener bool 1.6.0
-XX:StarvationMonitorInterval=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+StressLdcRewrite bool 1.6.0
-XX:+StressTieredRuntime bool 1.6.0
-XX:+SuppressFatalErrorMessage bool 1.6.0
-XX:SurvivorPadding=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:SurvivorRatio=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Sets the ratio of the survivor spaces (there are two) to the whole of the young generation.
-XX:SuspendRetryCount=<value> intx 1.6.0
-XX:SuspendRetryDelay=<value> intx 1.6.0
-XX:SyncFlags=<value> intx 1.6.0
-XX:SyncKnobs=<value> ccstr 1.6.0
-XX:SyncVerbose=<value> intx 1.6.0
-XX:+TaggedStackInterpreter bool 1.6.0
-XX:TargetPLABWastePct=<value> intx 1.6.0
-XX:TargetSurvivorOccupationRatio=<value> size_t 1.3.1
-XX:TargetSurvivorPadding=<value> size_t 1.3.1
-XX:TargetSurvivorRatio=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 This flag sets the desired percentage of the survivor space heap which must be used before objects are promoted to the old generation. For example, setting z to 90 would mean that 90% of the survivor space must be used before the young generation is considered full and objects are promoted to the old generation. This would allow objects to age more in the young generation before being tenured. The default value is 50.
-XX:TenuredGenerationSizeIncrement=<value> uintx 1.5.0 1.6.0
-XX:TenuredGenerationSizeSupplement=<value> uintx 1.5.0 1.6.0
-XX:TenuredGenerationSizeSupplementDecay=<value> uintx 1.5.0 1.6.0
-XX:ThreadPriorityPolicy=<value> intx 1.5.0 1.6.0
-XX:+ThreadPriorityVerbose bool 1.4.2 1.5.0 1.6.0
-XX:ThreadSafetyMargin=<value> intx 1.5.0 1.6.0
-XX:ThreadStackSize=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Thread Stack Size (in Kbytes) (0 means use default stack size) [Sparc: 512, Solaris Intel: 256, Sparc 64bit: 1024 all others 0]
-XX:ThresholdTolerance=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:Tier1BytecodeLimit=<value> intx 1.6.0
-XX:+Tier1UpdateMethodData bool 1.6.0
-XX:Tier2BackEdgeThreshold=<value> intx 1.6.0
-XX:Tier2CompileThreshold=<value> intx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+TieredCompilation bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:TLABAllocationWeight=<value> uintx 1.5.0 1.6.0
-XX:TLABFragmentationRatio=<value> uintx 1.4.0 1.4.1 1.4.2
-XX:TLABRefillWasteFraction=<value> uintx 1.5.0 1.6.0
-XX:TLABSize=<value> uintx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Set the size of the TLAB in bytes.
-XX:+TLABStats bool 1.5.0 1.6.0
-XX:TLABThreadRatio=<value> uintx 1.4.0 1.4.1 1.4.2
-XX:TLABWasteIncrement=<value> uintx 1.5.0 1.6.0
-XX:TLABWasteTargetPercent=<value> uintx 1.5.0 1.6.0
-XX:TLEFragmentationRatio=<value> 1.3.1 1.6.0
-XX:TLESize=<value> size_t 1.3.1 1.6.0 Sets the size of the TLEs (in bytes)
-XX:TLEThreadRatio=<value> 1.3.1 1.6.0
-XX:+TraceBiasedLocking bool 1.6.0
-XX:+TraceClassLoading bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Display classes as they are loaded. Similar to -verbose:class
-XX:+TraceClassLoadingPreorder bool 1.4.2 1.5.0 1.6.0
-XX:+TraceClassResolution bool 1.4.2 1.5.0 1.6.0
-XX:+TraceClassUnloading bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Display classes as they are unloaded.
-XX:+TraceFailedVMTextSegmentRetarget bool 1.4.0 1.4.1
-XX:+TraceGen0Time bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+TraceGen1Time bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:TraceJVMTI=<value> ccstr 1.5.0 1.6.0
-XX:+TraceJVMTIObjectTagging bool 1.5.0 1.6.0
-XX:+TraceLoaderConstraints bool 1.6.0
-XX:+TraceMonitorInflation bool 1.6.0
-XX:+TraceParallelOldGCTasks bool 1.6.0
-XX:TraceRedefineClasses=<value> intx 1.6.0
-XX:TypeProfileFailureLimit=<value> intx 1.4.0 1.4.1 1.4.2 1.5.0
-XX:TypeProfileMinimumRatio=<value> intx 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:Ultra3PrefetchCopyIntervalInBytes=<value> 1.3.1
-XX:Ultra3PrefetchScanIntervalInBytes=<value> 1.3.1
-XX:Ultra3PrefetchScavengeFieldsAhead=<value> 1.3.1
-XX:UnguardOnExecutionViolation=<value> intx 1.6.0
-XX:+UnlockDiagnosticVMOptions bool 1.4.2 1.5.0 1.6.0
-XX:+Use486InstrsOnly bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UseAdaptiveGCBoundary bool 1.5.0 1.6.0
-XX:+UseAdaptiveGenerationSizePolicyAtMajorCollection bool 1.5.0 1.6.0
-XX:+UseAdaptiveGenerationSizePolicyAtMinorCollection bool 1.5.0 1.6.0
-XX:+UseAdaptiveNUMAChunkSizing bool 1.6.0
-XX:+UseAdaptiveSizeDecayMajorGCCost bool 1.6.0
-XX:+UseAdaptiveSizePolicy bool 1.4.1 1.4.2 1.5.0 1.6.0 Automatically sizes the young generation and chooses an optimum survivor ratio to maximize performance.
-XX:+UseAdaptiveSizePolicyAggressiveFootprintGoal bool 1.5.0
-XX:+UseAdaptiveSizePolicyFootprintGoal bool 1.5.0 1.6.0
-XX:+UseAdaptiveSizePolicyWithSystemGC bool 1.5.0 1.6.0
-XX:+UseAdaptiveSizeScaleFootprintChangeByRatio bool 1.5.0
-XX:+UseAdaptiveSizeScaleThroughputChangeByRatio bool 1.5.0
-XX:+UseAltSigs bool 1.3.1 1.4.1 1.4.2 1.5.0 1.6.0 On the Solaris operating environment, the VM uses SIGUSR1 by default, which can sometimes conflict with applications that signal-chain SIGUSR1. -XX:+UseAltSigs will cause the VM to use signals other than SIGUSR1 and SIGUSR2 as the default.
-XX:+UseAutoGCSelectPolicy bool 1.6.0
-XX:+UseBiasedLocking bool 1.5.0 1.6.0 This option first appears in 1.5.0_06. Enables a technique for improving the performance of uncontended synchronization. Some applications with significant amounts of uncontended synchronization may attain significant speedups with this flag enabled.
-XX:+UseBoundThreads bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Bind user level threads (Solaris only)
-XX:+UseC2CallingConventions bool 1.3.1 1.4.0 1.4.1 1.4.2
-XX:+UseCMSBestFit bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UseCMSCollectionPassing bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UseCMSCompactAtFullCollection bool 1.4.1 1.4.2 1.5.0 1.6.0 Normally the concurrent low pause collector does not copy nor compact the live objects. A garbage collection is done without moving the live objects. If fragmentation becomes a problem, allocate a larger heap. In 1.4.2 if fragmentation in the tenured generation becomes a problem, a compaction of the tenured generation will be done although not concurrently. In 1.4.1 that compaction will occur if the UseCMSCompactAtFullCollection option is turned on.
-XX:+UseCMSInitiatingOccupancyOnly bool 1.4.1 1.4.2 1.5.0 1.6.0 Used to enabled/disable the use of the anticipated promotions to start a concurrent collection set
-XX:+UseCompiler bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UseCompilerSafepoints bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UseConcMarkSweepGC bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 This flag turns on concurrent garbage collection in the old generation. Available from J2SE1.4.1.
-XX:+UseDefaultStackSize bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0
-XX:+UseDepthFirstScavengeOrder bool 1.6.0
-XX:+UseExtendedFileIO bool 1.6.0
-XX:+UseFastAccessorMethods bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UseFastEmptyMethods bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UseFastJNIAccessors bool 1.5.0 1.6.0
-XX:+UseFixedFrameSize bool 1.4.0 1.4.1 1.4.2
-XX:+UseForcedSuspension bool 1.4.1 1.4.2
-XX:+UseGCOverheadLimit bool 1.6.0
-XX:+UseGCTaskAffinity bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UseHeavyMonitors bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UseInterpreter bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UseISM bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Intimate Shared Memory
-XX:+UseLargePages bool 1.6.0
-XX:+UseLargePageSize bool 1.4.0 1.4.1
-XX:+UseLoopCounter bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UseLWPSynchronization bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Use LWP-based instead of thread based synchronization. Default is on in J2SE 1.4.0 or higher. (Solaris only)
-XX:+UseMaximumCompactionOnSystemGC bool 1.6.0
-XX:+UseMembar bool 1.6.0
-XX:+UseMPSS bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Use Multiple Page Size Support (Solaris 9 only) w/4mb pages for the heap. Do not use with ISM as this replaces the need for ISM.
-XX:+UseNewCode bool 1.6.0
-XX:+UseNewCode2 bool 1.6.0
-XX:+UseNewCode3 bool 1.6.0
-XX:+UseNewFeature1 bool 1.4.2 1.5.0
-XX:+UseNewFeature2 bool 1.4.2 1.5.0
-XX:+UseNewFeature3 bool 1.4.2 1.5.0
-XX:+UseNewFeature4 bool 1.4.2 1.5.0
-XX:+UseNiagaraInstrs bool 1.6.0
-XX:+UseOnStackReplacement bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UseOversizedCarHandling bool 1.4.0 1.4.1 1.4.2 1.5.0 O
-XX:+UseParallelDensePrefixUpdate bool 1.6.0
-XX:+UseParallelGC bool 1.4.1 1.4.2 1.5.0 1.6.0 This flag also turns on parallel garbage collection policy in the young generation; however, it does not work with the CMS collector in the old generation. It is more suitable for enterprise applications which can benefit from very large young generation heaps.
-XX:+UseParallelOldGC bool 1.6.0
-XX:+UseParallelOldGCCompacting bool 1.6.0
-XX:+UseParallelOldGCDensePrefix bool 1.6.0
-XX:+UseParallelScavenge bool 1.3.1
-XX:+UseParNewGC bool 1.4.1 1.4.2 1.5.0 1.6.0 This flag turns on parallel garbage collection in the young generation. It can be enabled together with the CMS collector in the old generation. Available in J2SE1.4.1 on an experimental basis.
-XX:+UsePerfData bool 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UsePermISM bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UsePSAdaptiveSurvivorSizePolicy bool 1.5.0 1.6.0
-XX:+UseRecompilation bool 1.3.1
-XX:+UseResizableSurvivorSpaces bool 1.3.1
-XX:+UseSerialGC bool 1.5.0 1.6.0 Enable the “legacy” serial garbage collector. The 1.5+ collectors are generally more robust now, and this garbage collector will not be supported much longer
-XX:+UseSharedSpaces bool 1.5.0 1.6.0
-XX:+UseSignalChaining bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UseSpecLock bool 1.3.1 1.4.0
-XX:+UseSpinning bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UseSplitVerifier bool 1.5.0 1.6.0
-XX:UseSSE=<value> intx 1.5.0 1.6.0 From Bug 5108893, The option probably indicates whether the compiler should make use of the processor’s SSEi instructions and is only available in server VMs.
-XX:+UseStoreImmI16 bool 1.6.0
-XX:+UseSuspendResumeThreadLists bool 1.4.2 1.5.0 1.6.0
-XX:+UseThreadPriorities bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Use native thread priorities
-XX:+UseTLAB bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0 Turn off thread-local object allocation
-XX:+UseTLE bool 1.3.1 1.6.0 Turn off thread local portions of the heap in the young generation
-XX:+UseTrainGC bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 O
-XX:+UseTypeProfile bool 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+UseUnsupportedDeprecatedJVMPI bool 1.6.0
-XX:+UseVectoredExceptions bool 1.6.0
-XX:+UseVMInterruptibleIO bool 1.6.0
-XX:ValueMapBucketInitialSize=<value> intx 1.4.0
-XX:ValueMapBucketMaxSize=<value> intx 1.4.0
-XX:ValueMapMaxSize=<value> intx 1.4.0
-XX:+VerifyAfterGC bool 1.6.0
-XX:+VerifyAfterScavenge bool
-XX:+VerifyBeforeExit bool 1.6.0
-XX:+VerifyBeforeGC bool 1.6.0
-XX:+VerifyBeforeIteration bool 1.5.0 1.6.0
-XX:+VerifyBeforeScavenge bool
-XX:+VerifyDuringGC bool 1.6.0
-XX:VerifyGCLevel=<value> intx 1.6.0
-XX:+VerifyMergedCPBytecodes bool 1.6.0
-XX:+VerifyObjectStartArray bool 1.6.0
-XX:+VerifyRememberedSets bool 1.6.0
-XX:+VerifyTLAB bool 1.4.2
-XX:+VerifyTLE bool
-XX:+VMThreadHintNoPreempt bool 1.4.2 1.5.0 1.6.0
-XX:VMThreadPriority=<value> intx 1.4.2 1.5.0 1.6.0
-XX:VMThreadStackSize=<value> intx 1.3.1 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:WorkAroundNPTLTimedWaitHang=<value> intx 1.6.0
-XX:YoungGenerationSizeIncrement=<value> uintx 1.5.0 1.6.0
-XX:YoungGenerationSizeSupplement=<value> uintx 1.5.0 1.6.0
-XX:YoungGenerationSizeSupplementDecay=<value> uintx 1.5.0 1.6.0
-XX:YoungPLABSize=<value> uintx 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+ZeroTLAB bool 1.4.0 1.4.1 1.4.2 1.5.0 1.6.0
-XX:+ZeroTLE bool 1.3.1

References:

 

  1. JVM Performance Documentation
  2. Analyzing Stack Traces
  3. Turbo-charging Java HotSpot Virtual Machine
  4. Java HotSpot VM Options
  5. Tuning Garbage Collection with the 1.3.1 Java Virtual Machine
  6. Java and Solaris Threading
  7. Frequently Asked Questions about Garbage Collection in the HotspotTM JavaTM Virtual Machine
  8. Tuning Garbage Collection with the 1.4.2 Java[tm] Virtual Machine
  9. Improving Java Application Performance and Scalability by Reducing Garbage Collection Times and Sizing Memory Using JDK 1.4.1
  10. Big Heaps and Intimate Shared Memory (ISM)
  11. A Test of Java Virtual Machine Performance
  12. HotSpot Performance Tuning
  13. The Java HotSpot Virtual Machine Client Compiler
  14. Java Developer Connection Bug Parade
  15. SDK release notes 1.3.1.12
  16. JavaTM Virtual Machine Profiler Interface (JVMPI)
  17. Garbage Collector Ergonomics
  18. Java Developer Connection Forums
  19. Frequently Asked Questions About the Java HotSpot VM
  20. Java HotSpotTM Server VM Version 2.0 for Win32 Platforms
  21. J2SE 1.5 in a Nutshell
  22. 1.4.2 HotSpot Instrumentation
  23. HotSpot Technology Tools and HP-Specific Features
  24. Diagnosing a Garbage Collection problem
  25. GC Portal
  26. HotSpot Performance Tuning
  27. Java HotSpot VM Options (docs.sun.com)
  28. Java Tuning White Paper

发表评论