findutils support in 1.7.0

[Corinna Vinschen]

On Apr 28 11:05, Lev Bishop wrote:
> On Mon, Apr 28, 2008 at 10:57 AM, Corinna Vinschen wrote:
> 
> >  However, it's still strange that the path length (number of subdirs?)
> >  goes quadratic into the timing behaviour of a core OS routine to access
> >  a file.  The potential regression function evaluated by OpenOffice Calc
> >  is
> >
> >             5/3
> >   pathlength
> >   ------------- usecs
> >       100
> >
> >  and it appears to be a bit forgiving.
> 
> Using your "spreadsheet" from above, I get a very good fit using pure quadratic:
> 0.000558 pathlength^2 usec
> the residuals are ~100 usec and appear random.

Actually I used another table of values from a run of 1369 iterations
and a pathlength of up to 12338 chars. 

I ran the test on XP SP3.  And here I have... *gasp* ... positive news.

While I can reproduce the same almost quadratic behaviour on XP SP2, I
*can't* on Vista and 2008.  The behaviour appears to be rather linear,
even though it's hard to tell.  Anyway, Vista as well as 2008 are much
faster.  For the max tested pathlength of 12K chars, XP needs about
80000 usecs, Vista/2008 needs about 1200 usecs.  So there's actually
some hope in newer Windows versions.  For all of you who want to see the
actual timings, I appended a CSV file with what I measured.  All tests
were running on identical (virtual) hardware.


Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          cygwin AT cygwin DOT com
Red Hat
-------------- next part --------------
A non-text attachment was scrubbed...
Name: foo-values.csv
Type: text/x-comma-separated-values
Size: 33860 bytes
Desc: not available
URL: <http://cygwin.com/pipermail/cygwin-developers/attachments/20080428/2473fd35/attachment.bin>