26.3.6 Oddity

[Luke Jordan]

Hey friends--

 

I built out the solutions for this section on comparing sort and quick-sort, including the version of quick-sort that switches to sort when the list reaches a given length.  But in all the tests I've run for randomly generated lists, quick-sort is the fastest (according to time).  The documentation for time says something about "accuracy may vary by platform".  I'm running these on WinXP.  Based on the text, my expectation is that sort-quick-sort is faster than quick-sort for long lists.  But in my tests quick-sort has been the fastest in all cases.  This is in tests from lists of length 5 to 15,000.  I'm using 750 for a "switch-to-sort" threshold because nothing I tested that was much shorter than that would register > 0 milliseconds per time.

 

Has anyone else encountered this?  I've done the tests using languages from Intermediate through "full racket" without much difference.