🚀 go-pugleaf

    From: Daniel Finster <DF@sdf.lonestar.org>
    Date: Tue, 16 Jul 96 05:15 CDT
    
    FOO is a medium-sized (4925 element) list of strings, containing 1362 duplicates.

'string=' is a pretty slow function.  How long are the strings, and
where do the differences start to show up?

I think that the Genera implementation of 'remove-duplicates' has
complexity O( n(n-1) ), which isn't great.

Still, 932 seconds seems like an awfully long time.  I notice that the
GC flipped, too.  Who is consing?
     
        Evaluation of (REMOVE-DUPLICATES FOO :TEST #'STRING=) took 932.519780 seconds of elapsed time including:
          13.231 seconds processing sequence breaks,
          7.919 seconds in the storage system (including 4.649 seconds waiting for pages):
            6.749 seconds processing 7079 page faults including 216 fetches,
            0.804 seconds in creating and destroying pages, and
            0.366 seconds in miscellaneous storage system tasks.
        The garbage collector has flipped; so no consing was measured.
    
    Whats going on here?
    
    I'm running Genera 8.1.1 on a 3640, 7.5MW physical.