Forum Discussion

tlaford's avatar
tlaford
Occasional Contributor
14 years ago

Is TestComplete v8.70 slower than v8.60?

We just upgraded from 8.60 to 8.70 and when we ran our tests they are taking longer to complete.  For example:

   -a group of 2 tests that took 16 minutes to complete under v8.60 is taking more than 30 minutes to complete using v8.70

   -a group of 25 tests that took 4.5 hours to complete under v8.60 is taking about 8 hours to complete using v8.70



Has anyone else noticed this?



Does anyone have any idea on how to rectify this problem?



Could it be a problem with our converted project or a v8.70 configuration issue?

7 Replies

  • Hi Thomas,


    We are not aware of such a problem with TestComplete 8.70.

    The first thing that I would suggest is to check out whether there is some plug-in that causes the slowdown. You can disable all plug-ins you do not need for your testing and see whether this improves the performance. Here is the list of the plug-ins you can try disabling:


    .NET Open Application Support

    Adobe Flex Controls Support

    Delphi and C++ Builder Open Applications Support

    Firefox Support Plug-in

    Java Open Applications Support

    MSAA Open Applications

    Microsoft Visual C++ Open Applications Support

    QT Library Support Plug-in

    Silverlight Open Application Support

    UI Automation Support Plug-in

    Visual Basic Open Applications


    NOTE: when you disable the plug-ins, TestComplete will warn you that there are some other plug-ins that will not work. Click Yes in the dialog that asks whether you want to disable those plug-ins, as well.


    If this does not help, please send us your TestComplete project along with the results of its execution by versions 8.60 and 8.70 of the tool.

  • tlaford's avatar
    tlaford
    Occasional Contributor
    Hello Irene, than you for the reply.



    At the moment we have reverted to 8.60 in order to restore our ability to test in a timely manner.



    When time permits I will try your suggestion of disabling the plugins.



    I did a little more research:  I exported an 8.60 log and an 8.70 log and loaded these XML files into EXCEL.  Once in EXCEL I was able to calculate the delta time between each log entry (Error, Warning, Messahe, Event, Checkpoint) and then compare the 8.60 deltas with the 8.70 deltas.  What I noticed was that there was no single call that was responsible for the delay.  Most of the time deltas were between 1 second and 30 seconds longer in the 8.70 log.
  • Hi Thomas,


    It looks like it takes TC 8.70 more time to access your application's objects. Please describe the application under test and the test environment:


    - what programming language is used to develop the application?

    - what operating system is installed on the test machine?

    - do you use any specific Open Application techniques (MSAA, Text Recognition, UI Automation) to interact with your application's controls?


    When we get your answers, we will try to reproduce the performance slowdown here.

  • tlaford's avatar
    tlaford
    Occasional Contributor
    For the time being we have rolled back to version 8.60 in order to complete a delivery.  When time permits and we step up to 8.70 I will see what I can do.
  • tlaford's avatar
    tlaford
    Occasional Contributor
    After rolling back to TestComplete 8.60 we tried disabling the plugins as suggested.  We ended up disabling these plugins:

        Adobe Flex Controls Support

        Delphi and C++ Builder Open Applications Support

        Firefox Support Plug-in

        Java Open Applications Support

        Microsoft Visual C++ Open Applications Support

        QT Library Support Plug-in

        Silverlight Open Application Support

        UI Automation Support Plug-in

        Visual Basic Open Applications

    We needed to keep these plugins active

        .NET Open Application Support

        MSAA Open Applications



    Doing this on v8.60 reduced our total execution time (of 200 test scripts) from 23 hours to 16 hours!