Board index » delphi » floating point error and windows 98

floating point error and windows 98

I have a program that calculates a tridiagonal matrix.  I can set the
number of elements in the one dimensional arrays to over 10,000 using
Visual C++ at my university.  However, if I run the same code on Borland
C++ or Borland Builder 3 using Windows 98, I get floating point overflow
error with just 120 elements in my arrays. I am using double precision.
Float produces worst results.  I can only use 30 element arrays.  Is
this a problem with Borland and Windows 98? I would appreciate any
help.  I have a project due.
 

Re:floating point error and windows 98


Quote
On Thu, 22 Feb 2001 22:13:11 -0600, Doug Jordan <do...@pdq.net> wrote:
>I have a program that calculates a tridiagonal matrix.  I can set the
>number of elements in the one dimensional arrays to over 10,000 using
>Visual C++ at my university.  However, if I run the same code on Borland
>C++ or Borland Builder 3 using Windows 98, I get floating point overflow
>error with just 120 elements in my arrays. I am using double precision.
>Float produces worst results.  I can only use 30 element arrays.  Is
>this a problem with Borland and Windows 98? I would appreciate any
>help.  I have a project due.

Try changing the type to long double.

Jack Klein
--
Home: http://jackklein.home.att.net

Other Threads