What performance impact should one expect when setting UsePerfData=false?

David Dabbs dmdabbs at gmail.com
Sun Sep 12 04:24:44 PDT 2010


What performance impact (improvement) should one expect when setting

Also, I noticed that PerfDataSaveFile silently overrides PerfDataSaveToFile.

While anyone using this probably knows what they're doing, should
PrintFlagsFinal make this explicit? 

  /* performance data collection */

  product(bool, UsePerfData, true,            
          "Flag to disable jvmstat instrumentation for performance testing"
          "and problem isolation purposes.")  
  product(bool, PerfDataSaveToFile, false,    
          "Save PerfData memory to hsperfdata_<pid> file on exit")    
  product(ccstr, PerfDataSaveFile, NULL,      
          "Save PerfData memory to the specified absolute pathname,"  
           "%p in the file name if present will be replaced by pid")  

  // NOTE: If user specifies PerfDataSaveFile, it will save the performance
  // to the specified file name no matter whether PerfDataSaveToFile is
  // or not. In other word, -XX:PerfDataSaveFile=.. overrides flag
  // -XX:+PerfDataSaveToFile.

  product(intx, PerfDataSamplingInterval, 50 /*ms*/,     
          "Data sampling interval in milliseconds")      
  product(bool, PerfDisableSharedMem, false,  
          "Store performance data in standard memory")   
  product(intx, PerfDataMemorySize, 32*K,     
          "Size of performance data memory region. Will be rounded "  
          "up to a multiple of the native os page size.")
  product(intx, PerfMaxStringConstLength, 1024,          
          "Maximum PerfStringConstant string length before truncation")
  product(bool, PerfAllowAtExitRegistration, false,      
          "Allow registration of atexit() methods")      
  product(bool, PerfBypassFileSystemCheck, false,        
          "Bypass Win32 file system criteria checks (Windows Only)")  

More information about the hotspot-dev mailing list