| double precision complex #s [message #3960] |
Thu, 13 April 1995 00:00  |
psharer
Messages: 1
Registered: April 1995
|
Junior Member |
|
|
I am using the complex declaration for an array of complex numbers,
but am running into a problem with floating point overflows in a
polynomial calculation. Does anyone know of a way to create a complex
variable with double precision?
Thanks for your help,
Peter Sharer
|
|
|
|
| Re: double precision complex #s [message #4041 is a reply to message #3960] |
Mon, 17 April 1995 00:00  |
agraps
Messages: 35
Registered: September 1994
|
Member |
|
|
thompson@orpheus.nascom.nasa.gov (William Thompson) writes:
>> (There are more zeros than what's being printed here, but you get the
>> idea.)
> Uhhh, sorry to disappoint you, but simply taking a single precision number and
> converting it to double precision does not mean that it has double precision
> accuracy.
> Fear not, IDL v4.0 will have a DCOMPLEX data type.
> Bill Thompson
Yes, of course, but that's not what I was saying. Since I didn't know
where Peter's use of complex numbers were coming in for his calculation,
I was just offering a way for him to manipulate the real and imaginary
components in double precision separately (It'd be better to avoid the
complex stuff altogether.).
Amara
--
************************************************************ ***************
Amara Graps email: agraps@netcom.com
Computational Physicist vita: finger agraps@sunshine.arc.nasa.gov
Intergalactic Reality URL: http://www.best.com/~agraps/agraps.html
************************************************************ ****************
"I don't make jokes- I just watch the government and report the facts."
-- Will Rogers
|
|
|
|
| Re: double precision complex #s [message #4042 is a reply to message #3960] |
Mon, 17 April 1995 00:00 |
thompson
Messages: 584
Registered: August 1991
|
Senior Member |
|
|
agraps@netcom.com (Amara Graps) writes:
> psharer@eos.arc.nasa.gov (Peter J. Sharer) writes:
>> I am using the complex declaration for an array of complex numbers,
>> but am running into a problem with floating point overflows in a
>> polynomial calculation. Does anyone know of a way to create a complex
>> variable with double precision?
>> Thanks for your help,
>> Peter Sharer
> Peter,
> Try handling the real and imaginary parts this way (it's
> convoluted, but it works :) )
> IDL> a = [1,2,3]
> IDL> b = [4,5,6]
> IDL> c = complex(a,b)
> IDL> print, c
> ( 1.00000, 4.00000)( 2.00000, 5.00000)( 3.00000,
> 6.00000)
> To get double precision reals, use some of the rules of complex algebra,
> IDL> print, double(abs((c+conj(c))/2.0))
> 1.0000000 2.0000000 3.0000000
> To get double precision imaginary, use idl's imaginary function,
> IDL> print, double(imaginary(c))
> 4.0000000 5.0000000 6.0000000
> (There are more zeros than what's being printed here, but you get the
> idea.)
Uhhh, sorry to disappoint you, but simply taking a single precision number and
converting it to double precision does not mean that it has double precision
accuracy.
Fear not, IDL v4.0 will have a DCOMPLEX data type.
Bill Thompson
|
|
|
|
| Re: double precision complex #s [message #4045 is a reply to message #3960] |
Sun, 16 April 1995 00:00 |
agraps
Messages: 35
Registered: September 1994
|
Member |
|
|
psharer@eos.arc.nasa.gov (Peter J. Sharer) writes:
> I am using the complex declaration for an array of complex numbers,
> but am running into a problem with floating point overflows in a
> polynomial calculation. Does anyone know of a way to create a complex
> variable with double precision?
> Thanks for your help,
> Peter Sharer
Peter,
Try handling the real and imaginary parts this way (it's
convoluted, but it works :) )
IDL> a = [1,2,3]
IDL> b = [4,5,6]
IDL> c = complex(a,b)
IDL> print, c
( 1.00000, 4.00000)( 2.00000, 5.00000)( 3.00000,
6.00000)
To get double precision reals, use some of the rules of complex algebra,
IDL> print, double(abs((c+conj(c))/2.0))
1.0000000 2.0000000 3.0000000
To get double precision imaginary, use idl's imaginary function,
IDL> print, double(imaginary(c))
4.0000000 5.0000000 6.0000000
(There are more zeros than what's being printed here, but you get the
idea.)
Amara
--
************************************************************ ***************
Amara Graps email: agraps@netcom.com
Computational Physicist vita: finger agraps@sunshine.arc.nasa.gov
Intergalactic Reality URL: http://www.best.com/~agraps/agraps.html
************************************************************ ****************
"The map is not the territory." --Alfred Korzybski
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
