Re: Ron Knapp's time function
- To: mathgroup at smc.vnet.net
- Subject: [mg24612] Re: Ron Knapp's time function
- From: Robert Knapp <rknapp at wolfram.com>
- Date: Fri, 28 Jul 2000 17:23:48 -0400 (EDT)
- Organization: Wolfram Research, Inc.
- References: <8kjehf$e6o@smc.vnet.net>
- Sender: owner-wri-mathgroup at wolfram.com
fkampas wrote: > > Does anybody have the Mathematica code > for the generalized timing function that Ron > Knapp presented at the last developer's > conference? It isn't given in his notebooks > in the conference section on the Wolfram > web site. I tried posting this before, but it must have been lost: There is a more complete archive of talk notebooks I have used on my personal web page: http://members.wri.com/rknapp/Talks .... and therein you will find http://members.wri.com/rknapp/Talks/DeveloperConference99/time.nb which contains the function you seek. The time function is also used and included in my most recent Mathematica Journal article, which you can find pointers to through the main Wolfram research site www.wolfram.com ------------- Or simpler yet, here is the code SetAttributes[time, HoldAll]; time[expr_] := Block[{Second = 1}, Module[{t = First[Timing[expr]], tries = 1}, While[t < 1., tries *= 2; t = First[Timing[Do[expr, {tries}];]]]; t/tries ] ] ] The intent of this function is to allow you to compare timings of operations which may take only fractions of a second by using repetition to smooth over the timing granularity. It also suppresses the command output which, for people like me who are forgetful, avoids accidental huge outputs when comparing timings.