No announcement yet.

Repeating the same test - % variation in results

  • Filter
  • Time
  • Show
Clear All
new posts

  • Repeating the same test - % variation in results

    Friends, somebody did the Ptest on the same PC, for three times ?
    if YES, how much difference u found ? may be less than 1 percent ?

  • #2
    The difference between multiple test runs is generally greater than 1%. But should be less than 5% (except maybe for the disk test which can exhibit a wider range of results)

    This is generally because

    1/ Your typically windows PC has between 30 and 80 background processes and services running at any time. From time to time these processes can start up and grab a bit of CPU time. Some times you can control these activities (e.g. not moving the mouse & turning off virus scanners) but other times you might have less control (e.g. The arrival of a TCP/IP packet on your network connection.

    2/ The state of the computer will and does change from one test run to the next. Typical changes would be changes in the content of the CPU and disk cache, changes in memory locations were data is accessed and changes in the position of files on the disk (inner / outer cylinder + file fragmentation).

    3/ Sometimes there is also CPU throttling issues to think about. Some CPU throttle down when hot and other (in potables) throttle up under load.

    4/ CD drives that may sometimes start from a stopped state or sometimes already have the disk spun up to operation speed.

    But by being careful and observant you can avoid some of these effects.