MathGroup Archive 2004

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Bug report for MathixExp applied to a matrix of more-than-machine-precision numbers.


Thanks David,

I am working in 5.0.0 myself. In fact, closing Mathematica,
and pasting and evaluating the troublesome input resulted
in the answer.

So it seems like this was some glitch in my code then.
Sorry for generating noise :)

Sasha

David Park wrote:
> Sasha,
> 
> It works in 5.0.0.
> 
> MatrixExp[mat]
> 
> {{0.629715926773139433128201219496689223889890000573,
>     0.043272133628995532550007419267088674906992865756, \
> -0.775619348662370882970279342923966405559596934692}, \
> {-0.758122129833913698997991421827919657171707362198,
>     0.252008191520039596335048190450292895463503812363, \
> -0.601450503086404894412320958240615636790215514137}, \
> {0.169436382823573626402281120713124219469209126192,
>     0.966757153507535933149169968280951493199873521568,
>     0.191499133986736366584649123028828807021013925819}}
> 
> David Park
> djmp at earthlink.net
> http://home.earthlink.net/~djmp/
> 
> From: Oleksandr Pavlyk [mailto:pavlyk at phys.psu.edu]
To: mathgroup at smc.vnet.net
> 
> Hi,
> 
> I am not sure on how to report a bug in Mathematica, so
> I would write to the forum, knowing that the list is
> modetated :)
> 
> In[1]:=
> $Version
> 
> Out[1]=
> "5.0 for Microsoft Windows (June 11, 2003)"
> 
> In[2]:=
> mat = {{0,
> 0.6151525466966055419759862386321433310005256421644620580451761964922`49.296
> 724682967415,
> 
> -0.7254275036630691476762861174949573251764090399344597452533905218986`49.29
> 6724682967415},
> 
> {-0.6151525466966055419759862386321433310005256421644620580451761964922`49.2
> 96724682967415,
> 0,
> -1.2037607176452254987349374498805075260847239397930119102841747928461`49.29
> 6724682967415},
> 
> {0.7254275036630691476762861174949573251764090399344597452533905218986`49.29
> 6724682967415,
> 
> 1.2037607176452254987349374498805075260847239397930119102841747928461`49.296
> 724682967415,
> 0}};
> 
> In[3]:=
> MatrixExp[mat]
> 
> generates  Divide::indet,  Divide::infy, \[Infinity]::"indet" errors
> before Windows reports a crash of MathKernel, and shuts it down
> prompting to report the incident to Microsoft.
> 
> Trying to reduce the accuracy helps to get the final result
> 
> In[4]:=
> MatrixExp[N[mat,5]]
> 
> Out[4]=
> {{0.6297160719436076796`3.7756975065769325,
> 0.0432721390753584633`2.6576010298827444,
>     -0.7756190709530262614`3.833969641768763},
> {-0.7581218549510174401`3.8696850920659425,
>     0.25200848487286692`3.4668879073986285,
> -0.6014503110645328147`3.693042910203886},
>    {0.1694363428295855821`3.1961977983341225,
> 0.9667568181352670783`3.8868215583540806,
>     0.1914994511208260971`3.1806736750156337}}
> 
> Best regards,
> Sasha
> 
> 

-- 
Office: 6H Osmond                    Web: http://www.pavlyk.com
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
     No trees were destroyed to send this mail,
           but a lot of electrons were terribly disturbed.


  • Prev by Date: RE: Bug report for MathixExp applied to a matrix of more-than-machine-precision numbers.
  • Next by Date: Symbolic computation with vector fields and tensors
  • Previous by thread: RE: Bug report for MathixExp applied to a matrix of more-than-machine-precision numbers.
  • Next by thread: NMinimize and WebServices