Roy Longbottom at Linkedin  Roy Longbottom's JavaDraw Off and On Line Benchmarks


Contents


General JavaDrawPC JavaDrawPC Results
Revised Version Results Javax Swing Timer New Versions Without Swing
Java 8 Off-Line Results On-Line Versions On-Line Results
More Timer Measurements


General

As you may well know, Java source code is compiled into platform independent .class files and these are executed using a system specific Run Time Environment (RTE). Some might be compiled as Applets to run online via HTML documents. The initial benchmarks were offline varieties complied via Windows Command Prompt (or Linux terminals) using the javac command, such as javac test.java, where test.class is run via the command java test.

Java RTE might already be installed or older or latest versions downloaded. Javac is provided in Java Development Kit (JDK), for which versions for various platforms can be downloaded and installed. The first of my latest benchmarks was compiled on a PC using 64-Bit Windows 7 and JDK 7. Then run using RTE 1.7. Initially, Java security would not allow execution and no settings were available to overcome this. Googling solved the problem. Using a Java Uninstall Tool got rid of other Java versions and Security Settings were then visible. The latter needed a reduced setting to enable execution.

Other readily available Windows and Linux PCs had RTE 1.6 installed and the JDK 7 program produced a failure to run message. At a later stage, 32-Bit RTE 1.7 was installed on one system an this ran the program (from 64-Bit JDK 7). JDK 6 was installed on another PC and class file from this ran on available Windows and Linux based systems, using both JRE 1.6 and 1.7. With Windows, JDK might need the path adding to Computer Properties, Advanced System Settings, Environmental Variables, System Variables, Path. In my case, the following was added:

C:\Program Files (x86)\Windows Live\Shared;C:\Program Files\Java\jdk1.7.0_02\bin

Benchmark .java source codes, both JDK 6 and 7 .class runtime programs and copies of images used, are in Java PC Benchmarks.zip. The latter also includes an offline Java version of the Whetstone Benchmark.

A revised version was produced, repeating the first tests to identify startup overheads. Both the early versions produced unacceptable slow speeds on a new PC, due to a timer issue. A new version was then produced, using different timing arrangements. Appropriate files for these are included in the zip file.



To Start

JavaDrawPC

JavaDrawPC is intended to run the same test functions as my JavaDraw.apk benchmark for Android devices, where details and results can be found in Android Graphics Benchmarks.htm. There are also speeds measured on a Raspberry Pi. The benchmark uses small to rather excessive simple objects to measure drawing performance in Frames Per Second (FPS). Five tests draw on a background of continuously changing colour shades.

  • Test 1 loads two PNG files, one bitmap moving left/right for each frame, the other circling. A revised version runs this twice, the second one avoiding startup overheads.

  • Plus Test 2 for JavaDraw.apk generates 2 SweepGradient multi-coloured circles moving towards the centre and back. The circles are loaded a PNG file for this version.

  • Plus Test 3 draws 200 random small circles in the middle of the screen.

  • Plus Test 4 draws 80 lines from the centre of each side to the opposite side, again with changing colours.

  • Plus Test 5 draws the same small random circles as Test 3 but with 4000, filling the screen.

  • Each test runs for approximately 10 seconds at window size 1200 x 720 pixels.

Example Test 4 Display

JavaDraw Screen Copy


To Start

Measured speeds are displayed in the Command Prompt or Terminal window and in JavaDraw.txt or JavaDraw7.txt log files. An example of the revised version is shown below. The benchmark identifies the Operating System and RTE used along with limited information for the CPU used. Variations of the latter are also provided.


   Java Drawing Benchmark, Aug 14 2014, 11:10:20
            Produced by javac 1.7.0_65

  Test                              Frames      FPS

  Display PNG Bitmap Twice Pass 1     4803   479.58
  Display PNG Bitmap Twice Pass 2     5746   573.74
  Plus 2 SweepGradient Circles        6149   613.92
  Plus 200 Random Small Circles       3839   383.33
  Plus 320 Long Lines                 3834   382.83
  Plus 4000 Random Small Circles       467    46.63

         Total Elapsed Time  60.1 seconds

  Operating System    Windows 7, Arch. amd64, Version 6.1
  Java Vendor         Oracle Corporation, Version  1.7.0_02
  AMD64 Family 16 Model 4 Stepping 2, AuthenticAMD, 4 CPUs


 Other Variations

  Operating System    Windows 8.1, Arch. amd64, Version 6.3
                      Windows Vista, Arch. x86, Version 6.0
                      Windows Vista, Arch. amd64, Version 6.0
                      Windows XP, Arch. x86, Version 5.1
                      Linux, Arch. i386, Version 2.6.35-25-generic
                      Linux, Arch. amd64, Version 2.6.35-24-generic
                      Linux, Arch. amd64, Version 3.13.0-24-generic

  Javac               Produced by javac 1.6.0_27
                      Produced by javac 1.8.0_25

  Java Vendor         Sun Microsystems Inc., Version  1.6.0_21
                      Sun Microsystems Inc., Version  1.6.0_45
                      Oracle Corporation,    Version  1.7.0_55
                      Oracle Corporation,    Version  1.7.0_65
                      Oracle Corporation, Version  1.8.0_25

  x86 Family 6 Model 15 Stepping 13, GenuineIntel, 2 CPUs
  Intel64 Family 6 Model 15 Stepping 6, GenuineIntel, 2 CPUs
  Intel64 Family 6 Model 62 Stepping 4, GenuineIntel, 8 CPUs
  x86 Family 6 Model 28 Stepping 10, GenuineIntel, 2 CPUs
  null, null CPUs - not detected Linux

   

To Start

JavaDrawPC Frames Per Second Results

Following are speeds obtained via class files produced by JDK 6 and JDK 7, using JRE 1.6 or JRE 1.7. An example of results of the Android version are also included to show that the graphics is relatively slow compared with a Netbook and Desktop PCs. Note that taking into account power efficiency, like FPS per watt, there might be little difference. For example, looking at power supply specifications, the tablet’s is rated at 10 watts, the netbook’s at 40 watts and those for the desktop PCs more than 300 watts.

The results suggest that running tests 1 and 2 can be faster than test 1 by itself. However, there appears to be a high startup overhead and the difference disappears on running test 1 for a long time. When the benchmark is running, more than one CPU core can be used at the same time. Using Perfmon logging on the quad core Phenom showed average CPU utilisation of 57%, meaning concurrent use of more than two cores.


  JDK Compiler                 6       7       6       6       6       6       7
  CPU                       Phen    Phen     C2D     C2D    Atom    Atom    Atom Nexus 7
  Cores x GHz              4x3.0   4x3.0   2x2.4   2x2.4    1.66    1.66    1.66   4x1.4
  OS                        Win7    Win7   Vista  Ubuntu      XP  Ubuntu      XP Android
  Graphics                nVidia  nVidia  nVidia  nVidia   Intel   Intel   Intel  nVidia
                          GTS250  GTS250  8600GT  8600GT  945GSE  945GSE  945GSE     ULP
  JRE                        1.7     1.7     1.6     1.6     1.6     1.6     1.7

  PNG Bitmaps              489.8   486.0   343.3   261.2    47.5    73.7    59.4    20.4
  + SweepGradient Circles  591.8   580.4   466.2   221.9    63.3    76.7    75.4    16.5
  + 200 Small Circles      382.7   377.5   269.9   228.0    60.6    74.8    54.2    14.5
  + 320 Long Lines         385.2   376.9   292.2   203.8    47.3    51.9    42.0    11.3
  + 4000 Small Circles      46.4    46.7    36.0   136.2    12.4    24.1     9.6     3.8

   

To Start


Revised Version FPS Results

Below are results of the revised version, o check for where the first PNG Bitmaps test can have high initial overheads.

On testing a new PC, with a 3.7 GHz Core i7-4820K and Windows 8.1, performance was disappointingly slow. Then running under Linux Ubuntu 14.04, much more appropriate speeds were obtained. It seems that there is a timing issue on this system, details being provided in the next section.

       
                                Benchmarks JavaDrawPC6 and JavaDrawPC7

  JDK Compiler                 6       7       6       7       6       7       6       6
  CPU                       Phen    Phen Core i7 Core i7 Core i7 Core i7  Core 2  Core 2
  Cores x GHz              4x3.0   4x3.0   4x3.7   4x3.7   4x3.7   4x3.7   2x2.4  2x1.83
  OS                        Win7    Win7  Win8.1  Win8.1  Ubuntu  Ubuntu   Vista   Vista
  Graphics                nVidia  nVidia  nVidia  nVidia  nVidia  nVidia  nVidia  nVidia
                          GTS250  GTS250  GTX650  GTX650  GTX650  GTX650  8600GT   8400M
  JRE                        1.7     1.7     1.7     1.7     1.7     1.7     1.6     1.6

  PNG Bitmaps 1              494     482      61      62     322     508     357     125
  PNG Bitmaps 2              614     572      64      64     610     603     445     123
  + SweepGradient Circle     643     614      64      64     646     598     455     126
  + 200 Small Circles        392     385      64      64     532     485     288     128
  + 320 Long Lines           393     382      64      64     404     412     288     119
  + 4000 Small Circles        47      47      57      56     116     112      36      18
   

To Start


Javax Swing Timer

The benchmark sets up a timer, that automatically repaints the screen as a new frame after a specified delay, as timer = new Timer(0, x), indicating zero sleeping time before redrawing. To identify if something was wrong with the timer, MainClass.java was found, at the source shown below, that just counts passes with no graphics activity. The program was modified to repetitively repeat counting with no sleeping, producing Frames (passes) Per Second (FPS) and the time for each pass in microseconds - see listing.

On the Phenom, following some initial slow frame speeds, the time settled down to mainly less than 70 microseconds, to produce 543 FPS after 50 frames. The Windows 8.1 based PC ran mainly at a constant speed, taking somewhat higher than 15 milliseconds for each pass, resulting in 63 FPS for all passes. Sleeping time appears to be controlled by what is said to be Windows timeslice granularity of 1000 ms / 64 or 15.625 milliseconds.


           New PC Win 8.1      Phenom PC Win 7

    Frames  FPS   This frame   FPS   This frame
                  microsecs          microsecs

         1  500       2603     166       6170
         2   44      42630     333        100
         3   57       7051     500        107
         4   58      15402     571         74
         5   60      15550     555        231
         6   60      15633     117      41586
         7   60      15483     134        817
         8   61      15396     106      22895
         9   61      15388     107       8374
        10   61      15370     117        867
        11   62      15321     129         41
        12   62      15418     141        217
        13   62      15390     152         88
        14   62      15353     162         73
     To
        45   63      15353     494         45
        46   63      15360     505         48
        47   63      14998     516         45
        48   63      15306     521        113
        49   63      15331     532         64
        50   63      15365     543         44


     MainClass.java

    // From http://www.java2s.com/Code/JavaAPI/javax.swing/Timerstop.htm
    
    import java.awt.event.ActionEvent;
    import java.awt.event.ActionListener;
    import javax.swing.JFrame;
    import javax.swing.Timer;
    
    class MainClass extends JFrame 
    {
      Timer timer;
      long startTime, startTimeF, runTime, fps, usecs;
      int counter;
    
      MainClass(String title) 
      {
        setDefaultCloseOperation(JFrame.EXIT_ON_CLOSE);
       
        ActionListener a = new ActionListener() 
        {
          public void actionPerformed(ActionEvent e) 
          {
            runTime = (System.nanoTime() - startTime) / 1000000;
            usecs = (System.nanoTime() - startTimeF) / 1000;
            fps = counter * 1000 / runTime;
            System.out.println(" Frames = " + counter + "  " + fps + 
                       " FPS This frame " + usecs + " microseconds");
            startTimeF = System.nanoTime();
    
            if (++counter > 50) 
            {
              timer.stop();
              System.exit(0);
            }
          }
        };
    
        timer = new Timer(0, a);
        startTime = System.nanoTime();
        startTimeF = System.nanoTime();
        counter = 1;
        timer.start();
        setSize(800, 600);
    //    pack();
        setVisible(true);
      }
    
      public static void main(String[] args) 
      {
        new MainClass("Timer Demo1");
      }
    }
   

To Start


New Versions Without Swing - FPS Speeds

This version was rearranged to avoid using Java Swing functions but with Java AWT Threads. The Core i7 + Windows 8.1 PC runs at high speed with this. As before, PCs using JRE 6 will not run the programs produced with Javac 1.7. Also, speeds via Linux were somewhat different to those using Windows.

On the Phenom, CPU utilisation was up to 20% lower than that during the Swing tests.


                                 Benchmarks JavaDrawA6PC and JavaDrawA7PC

  JDK Compiler                 6       7       6       7       6       7       6       6
  CPU                       Phen    Phen Core i7 Core i7 Core i7 Core i7  Core 2  Core 2
  Cores x GHz              4x3.0   4x3.0   4x3.7   4x3.7   4x3.7   4x3.7   2x2.4  2x1.83
  OS                        Win7    Win7  Win8.1  Win8.1  Ubuntu  Ubuntu   Vista   Vista
  Graphics                nVidia  nVidia  nVidia  nVidia  nVidia  nVidia  nVidia  nVidia
                          GTS250  GTS250  GTX650  GTX650  GTX650  GTX650  8600GT   8400M
  JRE                        1.7     1.7     1.7     1.7     1.7     1.7     1.6     1.6

  PNG Bitmaps 1              628     568    1107    1112     862    1001     426     194
  PNG Bitmaps 2              734     692    1173    1176     982    1011     516     206
  + SweepGradient Circle     755     703    1161    1167     454     464     519     188
  + 200 Small Circles        416     419     620     618     454     452     296     179
  + 320 Long Lines           412     420     607     606     364     348     298     164
  + 4000 Small Circles        41      45      59      59     254     249      34      18
   

To Start


Java 8 Off-Line FPS Speeds

All the following results are from the 3.7 GHz Core i7, via Ubuntu 14.04. The first revised and new versions’ results use JRE 1.7.0_65 (slightly different to previous via 1.7.0_55). The others are from using JRE 1.8, the last one being compiled by javac 8. It can be seen that JRE 1.8 performance is completely different to that using the earlier JREs. Performance monitor indicates that JREs use more than one core, but CPU utilisation is much higher with JRE 1.8 at 35% x 8 threads, compared with 20% x 8 with JRE 1.7.


  CPU                    --------------------------- Core i7 ---------------------------
  Cores x GHz            ---------------------------   4x3.7 ---------------------------
  OS                     ---------------------------  Ubuntu ---------------------------
  Graphics               ---------------------------  nVidia ---------------------------
                         ---------------------------  GTX650 ---------------------------

  Version                    --------- Revised ---------     ------- New Version -------
  JDK Compiler                 7       6       7       8       7       6       7       8
  JRE                        1.7     1.8     1.8     1.8     1.7     1.8     1.8     1.8

  PNG Bitmaps 1              490    1626    1400    1589     779    1926    1971    1920
  PNG Bitmaps 2              591    1577    1869    1790     979    1997    2032    2083
  + SweepGradient Circle     580    1835    1702    1804     453    1936    1923    2048
  + 200 Small Circles        458     919     933     813     403     973     909     962
  + 320 Long Lines           379     305     379     364     307     315     312     383
  + 4000 Small Circles       107      60      63      56     219      42      41      43
     

To Start


On-Line Versions

Because of the slow speed problem, on-line versions were produced to see if the same thing occurred, and the one with swing functions also produced slow speeds, this time on both the Phenom and the Core i7 PCs.

These programs can be executed via the following links, when Java Control Panel Security might need a Medium setting to allow the programs to run. The on-line versions display results at the end, but they can be saved via using standard screen or window copying functions, such Alt PrtScr.

Note that the benchmark is designed to measure speed and displays might have flashing and missing objects, particularly with the on-line versions.


 Swing Versions

 http://www.roylongbottom.org.uk/javaonline/Java6SDraw.htm
 http://www.roylongbottom.org.uk/javaonline/Java7SDraw.htm

 AWT Versions

 http://www.roylongbottom.org.uk/javaonline/Java6ADraw.htm
 http://www.roylongbottom.org.uk/javaonline/Java7ADraw.htm
   

To Start


On-Line FPS Results

The first results below are for the new version where some of the speeds via Windows were significantly slower than the off-line test results (see last test + 4000 Small Circles). For these tests on the Core 2 system, the JDK 7 version also ran. This indicates that the browsers use JRE 1.7, not JRE 1.6 in the file path for the off-line version.

The second set of results are for on-line versions using Swing Timer. Here, PCs using both Windows 7 and 8 had maximum speeds limited by timeslice granularity of around 15.6 milliseconds, with Linux providing 1 millisecond. Vista speeds are much lower than the version without Swing, possibly due to CPU utilisation that appeared to be 100% of two cores.

The last table is from using Google Chrome and Firefox browsers, where both had timeslice granularity issues, with the former achieving full speeds on refreshing the page, but mainly slow on the initial access.

Later additions are for the Core i7, from Firefox running under Ubuntu 14.04. This required the use of IcedTea JRE. In this case, Swing and No-Swing speeds are similar.


  On-Line No Swing          Internet Explorer            FireFox            I Explorer  

  JDK Compiler                 6       7       6       7       6       7       6       7
  CPU                       Phen    Phen Core i7 Core i7 Core i7 Core i7  Core 2  Core 2
  Cores x GHz              4x3.0   4x3.0   4x3.7   4x3.7   4x3.7   4x3.7   2x2.4   2x2.4
  OS                        Win7    Win7  Win8.1  Win8.1  Ubuntu  Ubuntu   Vista   Vista
  Graphics                nVidia  nVidia  nVidia  nVidia  nVidia  nVidia  nVidia  nVidia
                          GTS250  GTS250  GTX650  GTX650  GTX650  GTX650  8600GT  8600GT
  JRE                        1.7     1.7     1.7     1.7     1.7     1.7     1.6     1.7

  PNG Bitmaps 1              440     440     941     932     969     953     419     423
  PNG Bitmaps 2              443     441     941     937     942    1003     435     435
  + SweepGradient Circle     306     312     726     722     467     467     370     372
  + 200 Small Circles        201     204     382     380     449     439     179      93
  + 320 Long Lines            62      65     156     154     406     401      96      85
  + 4000 Small Circles        22      23      33      33     273     272     8.3     7.5

 
  On-Line Swing             Internet Explorer            FireFox            I Explorer

  JDK Compiler                 6       7       6       7       6       7       6       7
  CPU                       Phen    Phen Core i7 Core i7 Core i7 Core i7  Core 2  Core 2
  Cores x GHz              4x3.0   4x3.0   4x3.7   4x3.7   4x3.7   4x3.7   2x2.4   2x2.4
  OS                        Win7    Win7  Win8.1  Win8.1  Ubuntu  Ubuntu   Vista   Vista
  Graphics                nVidia  nVidia  nVidia  nVidia  nVidia  nVidia  nVidia  nVidia
                          GTS250  GTS250  GTX650  GTX650  GTX650  GTX650  8600GT  8600GT
  JRE                        1.7     1.7     1.7     1.7     1.7     1.7     1.6     1.7

  PNG Bitmaps 1               64      64      64      64    1002     984     247     215
  PNG Bitmaps 2               64      64      64      64    1022    1006     255     238
  + SweepGradient Circle      64      64      64      64     472     485     232     213
  + 200 Small Circles         64      64      64      64     467     474     146     132
  + 320 Long Lines            60      60      64      64     408     412      77      72
  + 4000 Small Circles        22      23      27      27     286     306       6       5

 
                         Chrome                          FireFox
  On-Line                 Swing    Swing   Swing NoSwing   Swing   Swing NoSwing
 
  JDK Compiler            6 run1  6 run2       7       7       6       7       7
  CPU                       Phen    Phen    Phen    Phen    Phen    Phen    Phen
  Cores x GHz              4x3.0   4x3.0   4x3.0   4x3.0   4x3.0   4x3.0   4x3.0
  OS                        Win7    Win7    Win7    Win7    Win7    Win7    Win7
  Graphics                nVidia  nVidia  nVidia  nVidia  nVidia  nVidia  nVidia
                          GTS250  GTS250  GTS250  GTS250  GTS250  GTS250  GTS250
  JRE                        1.7     1.7     1.7     1.7     1.7     1.7     1.7

  PNG Bitmaps 1              155     416     415     438      64      64     436
  PNG Bitmaps 2               64     419     420     442      64      64     439
  + SweepGradient Circle      64     298     299     311      64      64     303
  + 200 Small Circles         64     200     201     205      64      64     202
  + 320 Long Lines            60      63      64      63      58      60      61
  + 4000 Small Circles        23      23      23      23      21      21      21
  

To Start


More Timer Measurements

Utilities are available to measure timeslice granularity and Swing function timing. Source addresses and results are shown below.

Utility 1 ClockRes - This appears to explain why Window 7 and 8, with current resolutions of around 15.6 milliseconds, produce a maximum of 64 frames per second on JavaDraw tests.

Utility 2 Swing TimeResolution - Results are provided for Linux Ubuntu and Windows Vista, where delays are what might be expected, with Windows 7 and 8 producing some anomalies.

New Timer - On creating a new Timer, a delay is specified (assumed to be in milliseconds) that causes a Listener function to be executed at intervals dictated by the delay. The results show that windows 7 and 8, except the former with zero delay, provide Listener refreshes at 15.6 milliseconds, doubling to 31.2 seconds where 16 is specified.

sleep function - On Windows 7 and 8, this apparently produces the wronf delays with 10 and 20 millisecond parameters.

wait function - Windows 7 and 8 results are the same as from a new Timer, starting with a parameter of 1.

Other Timers Wrong - Java Thread.sleep(n) and C/C++ Sleep(n) function timings were measured and, at least for low values of n milliseconds, produced unexpected longer delays.


 ClockRes from 

 http://technet.microsoft.com/en-US/sysinternals 

 ClockRes v2.0 - View the system clock resolution
 Copyright (C) 2009 Mark Russinovich

 SysInternals - www.sysinternals.com

                             Win 7    Win 8    Vista
                             msecs    msecs    msecs

  Maximum timer interval:   15.600   15.625   15.600
  Minimum timer interval:    0.500    0.500    0.500
  Current timer interval:   15.600   15.625    1.000


  Swing timer milliseconds via 

  http://www.java2s.com/Code/Java/Swing-JFC/TimeResolution.htm 

  Timer calls an action listener at regular intervals specified by delay 

    Javac    1.7     1.6     1.7     1.6     1.7     1.6     1.6     1.6
    Java     1.7     1.7     1.7     1.7     1.7     1.7     1.6     1.6
     CPU Core i7 -----------------------  Phenom -------  Core 2a Core 2b
     OS  Win 8.1--------  Ubuntu 14 ----   Win 7 -------  Win Vista ----
    Bits      64      64      64      64      64      64      32      64

    new
    Timer  ------------------------- per delay -------------------------
    delay
       0   15.66   15.66    0.18    0.72    0.11    0.22    1.04    1.08
       1   15.63   15.63    1.20    1.34   15.60   15.60    1.01    1.04
       2   15.62   15.62    2.18    2.18   15.60   15.60    2.02    3.22
       3   15.62   15.62    3.18    3.18   15.60   15.60    3.00    3.23
       4   15.62   15.62    4.18    4.18   15.60   15.60    4.00    4.25
       5   15.63   15.63    5.18    5.18   15.60   15.60    5.00    5.43
       6   15.62   15.62    6.17    6.17   15.60   15.60    6.01    6.61
       7   15.62   15.62    7.17    7.18   15.60   15.60    7.01    7.97
       8   15.62   15.62    8.18    8.17   15.60   15.60    8.00    8.80
       9   15.62   15.62    9.17    9.17   15.59   15.59    9.00    9.09
      10   15.62   15.62   10.17   10.17   15.60   15.60   10.00   10.07
      11   15.62   15.62   11.17   11.17   15.60   15.60   11.00   11.04
      12   15.63   15.63   12.18   12.17   15.59   15.59   12.00   12.18
      13   15.62   15.62   13.17   13.17   15.59   15.59   13.00   13.09
      14   15.62   15.62   14.17   14.18   15.59   15.59   14.00   14.10
      15   15.62   15.62   15.17   15.17   15.59   15.59   15.00   15.00
      16   31.24   31.24   16.18   16.18   31.19   31.19   16.00   16.05
      17   31.24   31.24   17.17   17.17   31.19   31.19   17.00   17.28
      18   31.24   31.24   18.16   18.15   31.20   31.20   18.00   18.00
      19   31.25   31.25   19.15   19.15   31.19   31.19   19.00   19.00
      20   31.24   31.24   20.16   20.16   31.20   31.20   20.00   20.04
           XXXXX   XXXXX                   ?XXXX   ?XXXX


    sleep   ------------------------ per sleep -------------------------
     time
       0    0.00    0.00    0.00    0.00    0.00    0.00    0.00    0.00
       1    1.03    1.03    1.09    1.09    1.02    1.02    1.00    1.01
       2    2.13    2.15    2.08    2.09    2.03    2.00    2.00    2.04
       3    3.16    3.16    3.08    3.09    3.05    3.03    3.00    3.05
       4    4.22    4.19    4.08    4.09    4.06    4.00    4.00    4.04
       5    5.23    5.14    5.09    5.09    5.07    5.01    5.00    5.07
       6    6.21    6.19    6.08    6.08    6.08    5.99    5.99    6.07
       7    7.37    7.35    7.08    7.08    7.09    6.99    6.99    7.11
       8    8.16    8.24    8.08    8.08    8.06    7.99    7.99    8.08
       9    9.26    9.16    9.08    9.08    9.08    8.99    8.99    9.01
      10   15.61X  15.61X  10.08   10.08   15.61X  15.61X   9.99   10.09
      11   11.29   11.22   11.08   11.09   11.13   11.09   10.99   11.04
      12   12.20   12.23   12.08   12.08   12.08   12.05   11.99   12.00
      13   13.30   13.25   13.08   13.08   13.04   12.99   13.00   12.99
      14   14.34   14.52   14.08   14.08   14.44   13.99   14.00   14.01
      15   15.17   15.18   15.08   15.08   15.02   14.98   14.98   15.11
      16   16.15   16.24   16.08   16.08   16.08   15.98   15.98   16.08
      17   17.22   17.17   17.09   17.09   17.00   16.98   16.98   17.05
      18   18.11   18.31   18.07   18.07   18.02   17.98   17.98   18.00
      19   19.25   19.50   19.08   19.08   19.00   18.98   19.00   19.13
      20   31.22X  31.22X  20.08   20.08   31.22X  31.22X  19.98   19.98

     wait  -------------------------- per wait ------------------------- 
     time
       1   15.62   15.62    1.09    1.09   15.60   15.60    1.01    1.01
       2   15.62   15.62    2.09    2.09   15.60   15.60    2.00    2.04
       3   15.62   15.62    3.08    3.09   15.60   15.60    3.00    3.05
       4   15.62   15.62    4.08    4.09   15.60   15.60    4.00    4.06
       5   15.61   15.61    5.09    5.09   15.60   15.60    4.99    5.04
       6   15.61   15.61    6.08    6.09   15.59   15.59    5.99    6.05
       7   15.61   15.61    7.08    7.09   15.59   15.59    7.00    7.01
       8   15.61   15.61    8.08    8.09   15.59   15.59    7.99    8.02
       9   15.60   15.60    9.08    9.08   15.59   15.59    9.01    9.03
      10   15.61   15.61   10.08   10.08   15.59   15.59    9.99   10.03
      11   15.60   15.60   11.08   11.09   15.59   15.59   10.99   11.12
      12   15.61   15.60   12.08   12.08   15.58   15.58   12.06   12.02
      13   15.61   15.61   13.08   13.08   15.58   15.58   12.99   13.07
      14   15.61   15.59   14.08   14.08   15.58   15.58   13.99   14.07
      15   15.59   15.59   15.08   15.09   15.58   15.58   14.98   15.23
      16   31.23   31.23   16.08   16.08   31.18   31.18   15.98   16.08
      17   31.22   31.22   17.09   17.09   31.17   31.17   16.98   17.02
      18   31.22   31.22   18.07   18.07   31.18   31.18   17.98   18.05
      19   31.23   31.21   19.08   19.08   31.17   31.17   18.98   19.04
      20   31.22   31.22   20.08   20.10   31.18   31.18   19.98   20.14
           XXXXX   XXXXX                   XXXXX   XXXXX   
  

To Start


Roy Longbottom at Linkedin   Roy Longbottom January 2015

The Official Internet Home for my Benchmarks is via the link
Roy Longbottom's PC Benchmark Collection