Tuesday, March 03, 2009

Getting Results Takes Time

Who would have thought it would take an entire hour to get ten measurements on the time it takes to open a Word document? It takes time for Windows to boot, and then the software waits 200 seconds before attempting to open the file. I am using my DelayLaunch utility to wait for 200 seconds: there is a shortcut in the startup folder to launch it.
Then there is the time taken to note the result in a spreadsheet before rebooting and getting the next measurement. I allow 180 seconds (3 minutes) for Windows to boot up and settle down (?), and then wait another 20 seconds before doing a timed opening of the file.
Click on the top graphic to see more details of the result obtained. The VB program notes the time on the left, in minutes and seconds, and also seconds since midnight. Then it asks the OS to open the file, and displays the time on the right once this process has finished. Also, the opened document displays the current time, for checking purposes.
I am taking an average of only 10 results. Statistically 21 results should be used, but it just takes too long, and the FileAccessTimer program only gets 10 results, so there isn't much point in getting more. So far the results vary by ±15% which is higher than I thought it would be. That means any result could be 15% higher or 15% lower than the average of 10 results. It's quite a wide variation for what is supposed to be a repeatable sequence, but I have noticed the same variation when using FileAccessTimer to open all the files in the "Program Files" folder. Who says defrag benchmarking is an exact science? Oh well ...

Update: Friday 13th March 2009
: I have put together a "definitive" (for now) set of results and tests for Windows 5.1.2600 (i.e WinXP SP3 32-bit) and I'm in the process of getting results for Windows Disk Defragmenter. This should take a few more days.
I have also purchased a 64-bit version of Windows Vista, which has been paid for by the Google Adwords revenue for the last 9 months or so, thanks to all those who don't use IE7Pro's Ad Blocker or Firefox's AdBlock Plus. This blog is not a profit-making exercise, but a hobby.

No comments: