Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.soft-sys.math.mathematica > #5392

Re: count zeros in a number

From Richard Fateman <fateman@eecs.berkeley.edu>
Newsgroups comp.soft-sys.math.mathematica
Subject Re: count zeros in a number
Date 2011-10-05 08:08 +0000
Organization Steven M. Christensen and Associates, Inc and MathTensor, Inc.
Message-ID <j6h397$721$1@smc.vnet.net> (permalink)
References <201110020636.CAA28027@smc.vnet.net> <j6brvm$8om$1@smc.vnet.net> <201110040530.BAA20698@smc.vnet.net> <op.v2t544pxtgfoz2@bobbys-imac.local>

Show all headers | View raw


On 10/4/2011 9:44 AM, DrMajorBob wrote:
> There's nothing "esoteric" about IntegerDigits, Replace, or Repeated, 
> as in:
>
> x = 24^24*55^55;
> Replace[IntegerDigits@x, {__, zeroes : 0 ..} :>
>    Length@{zeroes}] // Timing
>
> {0.000185, 55}

My feeling is that these ARE esoteric.  A person just introduced to 
Mathematica would probably not encounter IntegerDigits, Repeated, or 
patterns other than the trivial one used in pseudo function definitions as
f[x_]:=   x+1.


>
> If we wanted brute-force arithmetic, we might use Fortran.
>
> Your suggested solutions are NOT greatly hindered by "the slow 
> implementation of looping constructs in Mathematica", on the other hand:
>
> (n = 0; While[GCD[x, 10^n] == 10^n, n++]; n - 1) // Timing
>
> {0.000029, 0}

huh?   I got  {0., 55}.   Perhaps my computer is faster or has a lower 
resolution timer, but we should both get 55.

>
> (n = 0; While[Mod[x, 10] == 0, (x = x/10; n++)]; n) // Timing
>
> {0.000018, 0}

I mention the slow implementation of looping constructs simply to warn 
people that if you can resolve your computation by simple composition of 
built-in functions e.g. F[G[H[x]]]   and don't rely on Mathematica
to efficiently execute loops written as While[] For[]  etc, you are 
probably going to run faster.  Let Mathematica's operation on compound 
objects like lists do the job.  Length[] is a lot faster than counting 
with a While etc.

In my timings, doing the operations 1000 times..
Do[.....,{1000}]//Timing

I found that my solutions were about 10X faster than the IntegerDigits 
one, on this particular value of x.

Fortran can't be used unless you manage arbitrary precision integers...

RJF



Back to comp.soft-sys.math.mathematica | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread


Thread

Re: count zeros in a number "Harvey P. Dale" <hpd1@nyu.edu> - 2011-10-03 08:32 +0000
  Re: count zeros in a number Richard Fateman <fateman@cs.berkeley.edu> - 2011-10-04 05:31 +0000
  Re: count zeros in a number DrMajorBob <btreat1@austin.rr.com> - 2011-10-05 08:05 +0000
  Re: count zeros in a number Richard Fateman <fateman@eecs.berkeley.edu> - 2011-10-05 08:08 +0000
  Re: count zeros in a number DrMajorBob <btreat1@austin.rr.com> - 2011-10-05 08:09 +0000
  Re: count zeros in a number DrMajorBob <btreat1@austin.rr.com> - 2011-10-05 08:11 +0000
  Re: count zeros in a number Richard Fateman <fateman@eecs.berkeley.edu> - 2011-10-05 08:12 +0000
  Re: count zeros in a number DrMajorBob <btreat1@austin.rr.com> - 2011-10-05 08:13 +0000

csiph-web