Re: Question on trailing zeros
- To: mathgroup at smc.vnet.net
- Subject: [mg40002] Re: Question on trailing zeros
- From: "Allan Hayes" <hay at haystack.demon.co.uk>
- Date: Fri, 14 Mar 2003 04:46:46 -0500 (EST)
- References: <b4pfgo$bus$1@smc.vnet.net>
- Sender: owner-wri-mathgroup at wolfram.com
Carlos,
NSolve[-3-6*g5+21*g5^2+28*g5^3==0,g5] computes a solution, sol say, and then
SetPrecision[sol, 64] adds zeros to the numbers in this in their binary form
to give numbers to precision 64 in decimal form. This is where the zeros
come from, but it is rather artificial here and has nothing to do with the
precision of the answer. To get an answer with precision 64 we must use the
form
NSolve[-3-6*g5+21*g5^2+28*g5^3==0,g5,64]
NSolve[-3 - 6*g5 + 21*g5^2 + 28*g5^3 == 0, g5, 64]
"Carlos Felippa" <carlos at colorado.edu> wrote in message
news:b4pfgo$bus$1 at smc.vnet.net...
> sol=SetPrecision[NSolve[-3-6*g5+21*g5^2+28*g5^3==0,g5],64];
> g5roots=g5/.sol; Print[g5roots];
>
> {-0.8540119518537006015890256094280630350112915039062500000000000000,
> -0.3059924679232961519659284022054634988307952880859375000000000000,
> 0.4100044197769967535549540116335265338420867919921875000000000000}
>
> What is the meaning of those zeros at the end of the display?
> Noise digits?
>