Re: More problems with SetPrecision[] and/or $MinPrecision,...
- To: mathgroup at smc.vnet.net
- Subject: [mg57025] Re: More problems with SetPrecision[] and/or $MinPrecision,...
- From: dh <dh at metrohm.ch>
- Date: Thu, 12 May 2005 02:34:30 -0400 (EDT)
- References: <d5sjpu$nph$1@smc.vnet.net>
- Sender: owner-wri-mathgroup at wolfram.com
Hi Terry,
MMA version: 5.1, Windows
in my case everything work as it should.
I get:
0.x10^-21
this is the correct answer for a value of zero that has an accuracy of
21 and (by definition) zero precision
Sincerely, Daniel
terryisnow at yahoo.com wrote:
> Dear Mathematica Guru's,
>
> Lately I seem to be finding a lot of SetPrecision[]/$MinPrecision bugs.
> This appears to be another one:
>
> In[1]:= $MinPrecision = 20;
>
> savedData1 = 669151541.9328941107875;
> savedData2 = 0.99960472897267246018765;
> savedData3 = Log[savedData1^savedData2/savedData1^savedData2];
>
> Print[Apply[Plus, savedData3 -
> Log[savedData1^savedData2/savedData1^savedData2]] ];
>
> x = 0.99960472897267245645;
> y = 669151541.932894110712;
> Do[
> z = SetPrecision[y,24]^x;
> ,{50000}];
>
> Print[Apply[Plus, savedData3 -
> Log[savedData1^savedData2/savedData1^savedData2]] ];
>
> The output looks like this on my Mathematica 3.0 system:
>
> 0
> 1.9586707534418188626 x 10^-30
>
> Obviously the question is why is the second Print statement not
> also returning a 0, after all the code that runs between the two
> Print statements neither changes any of the variables referenced
> by the two Print statements, nor even references any of the same
> variables.
>
> The problem maybe reproducible without the $MinPrecision but with
> this specific test case it is needed (there might however be other
> constants that can reproduce the problem without $MinPrecision).
>
> Can someone also run this on a more recent version of Mathematica
> to see if it reproduces? All help is very much appreciated!
>
> Terry
>