comp.lang.idl-pvwave archive: archive

Home » Public Forums » archive » IDL 8.2.2 released

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

Re: IDL 8.2.2 released [message #83169 is a reply to message #83056]

Fri, 08 February 2013 08:24

dain.cilke
Messages: 3
Registered: February 2013

Junior Member

On Thursday, February 7, 2013 3:55:14 PM UTC-7, bobnn...@gmail.com wrote:
> On Thursday, February 7, 2013 3:10:32 PM UTC-7, dain....@gmail.com wrote:
>
>> On Thursday, February 7, 2013 11:53:16 AM UTC-7, bobnn...@gmail.com wrote:
>
>>
>
>>> On Wednesday, February 6, 2013 2:23:04 PM UTC-7, bobnn...@gmail.com wrote:
>
>>
>
>>>
>
>>
>
>>>> I wonder if the very long existing bug where INTERPOLATE internally truncates all results to single precision even when double precision input is used is still present (note that it still returns the results in a double precision variable but half the precision is gone). It is in 8.1 so I have little hope it is gone in 8.2.2.
>
>>
>
>>>
>
>>
>
>>>>
>
>>
>
>>>
>
>>
>
>>>>
>
>>
>
>>>
>
>>
>
>>>>
>
>>
>
>>>
>
>>
>
>>>> Am I the only one that has been burned by this?
>
>>
>
>>>
>
>>
>
>>>
>
>>
>
>>>
>
>>
>
>>> Well, I guess I am the only one who cares. But in case people just didn't know what I meant here is an example. Note that this is a toy example. The point is that if you have a function which you expect to work in double precision and you call interpolate then your results are truncated to single with no indication. I know of no other IDL built in function that does this!
>
>>
>
>>>
>
>>
>
>>>
>
>>
>
>>>
>
>>
>
>>> ;; File interpolate_test.pro
>
>>
>
>>>
>
>>
>
>>>
>
>>
>
>>>
>
>>
>
>>> x = dindgen(6)
>
>>
>
>>>
>
>>
>
>>> y = dindgen(6)
>
>>
>
>>>
>
>>
>
>>> x_i = dindgen(5) + 0.1d
>
>>
>
>>>
>
>>
>
>>>
>
>>
>
>>>
>
>>
>
>>> print, 'INTERPOLATE: incorrect answer, accurate only to 8 digits'
>
>>
>
>>>
>
>>
>
>>> print,interpolate(y,x_i),FORMAT='(G20.16)'
>
>>
>
>>>
>
>>
>
>>>
>
>>
>
>>>
>
>>
>
>>> print, 'INTERPOL: correct answer, accurate to 16 digits'
>
>>
>
>>>
>
>>
>
>>> print,interpol(y,x,x_i),FORMAT='(G20.16)'
>
>>
>
>>>
>
>>
>
>>>
>
>>
>
>>>
>
>>
>
>>> end
>
>>
>
>>>
>
>>
>
>>>
>
>>
>
>>>
>
>>
>
>>> IDL> .run interolate_test.pro
>
>>
>
>>>
>
>>
>
>>> % Compiled module: $MAIN$.
>
>>
>
>>>
>
>>
>
>>> INTERPOLATE: incorrect answer, accurate only to 8 digits
>
>>
>
>>>
>
>>
>
>>> 0.1000000014901161
>
>>
>
>>>
>
>>
>
>>> 1.100000023841858
>
>>
>
>>>
>
>>
>
>>> 2.099999904632568
>
>>
>
>>>
>
>>
>
>>> 3.099999904632568
>
>>
>
>>>
>
>>
>
>>> 4.099999904632568
>
>>
>
>>>
>
>>
>
>>> INTERPOL: correct answer, accurate to 16 digits
>
>>
>
>>>
>
>>
>
>>> 0.1000000000000000
>
>>
>
>>>
>
>>
>
>>> 1.100000000000000
>
>>
>
>>>
>
>>
>
>>> 2.100000000000000
>
>>
>
>>>
>
>>
>
>>> 3.100000000000000
>
>>
>
>>>
>
>>
>
>>> 4.100000000000000
>
>>
>
>>
>
>>
>
>> Hi Bob,
>
>>
>
>> Thanks for bring this to our attention! This bug has been fixed. Now interpolate will perform all calculations in double precision.
>
>>
>
>>
>
>>
>
>> Cheers,
>
>>
>
>> Dain
>
>>
>
>> IDL Code Monkey
>
>>
>
>> ExelisVIS
>
>
>
> OK, that is great.
>
>
>
> However, I assume that single precision inputs will still be done in single precision (or at least returned in single precision). I dislike routines that arbitrarily change the precision.
>
>
>
> Bob

Hi Bob,

Single precision inputs will result in single precision outputs. However, the internals of the function will be in double precision.

Hope this helps,
Dain

Report message to a moderator

[Message index]

		IDL 8.2.2 released By: Mark Piper on Tue, 05 February 2013 09:39
		Re: IDL 8.2.2 released By: David Fanning on Wed, 06 February 2013 16:24
		Re: IDL 8.2.2 released By: Mark Piper on Thu, 07 February 2013 08:16
		Re: IDL 8.2.2 released By: Mark Piper on Wed, 06 February 2013 16:09
		Re: IDL 8.2.2 released By: tom.grydeland on Wed, 20 March 2013 02:58
		Re: IDL 8.2.2 released By: Mark Piper on Wed, 06 February 2013 16:00
		Re: IDL 8.2.2 released By: Matt[2] on Thu, 07 February 2013 08:20
		Re: IDL 8.2.2 released By: Mark Piper on Wed, 06 February 2013 15:23
		Re: IDL 8.2.2 released By: Bob[4] on Wed, 06 February 2013 13:23
		Re: IDL 8.2.2 released By: Bob[4] on Thu, 07 February 2013 10:53
		Re: IDL 8.2.2 released By: dain.cilke on Thu, 07 February 2013 14:07
		Re: IDL 8.2.2 released By: Louis Giglio on Wed, 06 February 2013 13:01
		Re: IDL 8.2.2 released By: Andy Sayer on Thu, 07 February 2013 06:22
		Re: IDL 8.2.2 released By: Fabzi on Wed, 06 February 2013 11:18
		Re: IDL 8.2.2 released By: David Fanning on Wed, 06 February 2013 10:24
		Re: IDL 8.2.2 released By: Mark Piper on Wed, 06 February 2013 09:20
		Re: IDL 8.2.2 released By: Mark Piper on Wed, 06 February 2013 09:18
		Re: IDL 8.2.2 released By: David Fanning on Wed, 06 February 2013 06:51
		Re: IDL 8.2.2 released By: Fabzi on Wed, 06 February 2013 06:37
		Re: IDL 8.2.2 released By: lecacheux.alain on Wed, 06 February 2013 04:29
		Re: IDL 8.2.2 released By: Fabzi on Wed, 06 February 2013 01:44
		Re: IDL 8.2.2 released By: wlandsman on Tue, 05 February 2013 23:29
		Re: IDL 8.2.2 released By: David Fanning on Tue, 05 February 2013 15:06
		Re: IDL 8.2.2 released By: wlandsman on Tue, 05 February 2013 14:28
		Re: IDL 8.2.2 released By: timothyja123 on Sat, 09 February 2013 02:44
		Re: IDL 8.2.2 released By: Matt Haffner on Fri, 08 February 2013 13:48
		Re: IDL 8.2.2 released By: Michael Galloy on Wed, 13 February 2013 16:19
		Re: IDL 8.2.2 released By: Mark Piper on Wed, 20 March 2013 14:10
		Re: IDL 8.2.2 released By: Fabzi on Fri, 08 February 2013 09:15
		Re: IDL 8.2.2 released By: Fabzi on Fri, 08 February 2013 09:04
		Re: IDL 8.2.2 released By: timothyja123 on Tue, 12 February 2013 16:43
		Re: IDL 8.2.2 released By: dain.cilke on Fri, 08 February 2013 08:24
		Re: IDL 8.2.2 released By: chris_torrence@NOSPAM on Fri, 08 February 2013 08:14
		Re: IDL 8.2.2 released By: Bob[4] on Thu, 07 February 2013 14:55
		Re: IDL 8.2.2 released By: dain.cilke on Thu, 07 February 2013 14:10
		Re: IDL 8.2.2 released By: Bob[4] on Thu, 07 February 2013 12:14
		Re: IDL 8.2.2 released By: David Fanning on Thu, 07 February 2013 11:15
		Re: IDL 8.2.2 released By: chris_torrence@NOSPAM on Thu, 07 February 2013 09:11
		Re: IDL 8.2.2 released By: Mark Piper on Thu, 07 February 2013 08:55
		Re: IDL 8.2.2 released By: David Fanning on Thu, 07 February 2013 08:21
		Re: IDL 8.2.2 released By: Mark Piper on Thu, 14 February 2013 08:23
		Re: IDL 8.2.2 released By: Michael Galloy on Wed, 13 February 2013 16:16
		Re: IDL 8.2.2 released By: timothyja123 on Wed, 13 February 2013 14:52
		Re: IDL 8.2.2 released By: Fabzi on Wed, 13 February 2013 02:04
		Re: IDL 8.2.2 released By: lecacheux.alain on Wed, 20 March 2013 09:07
		Re: IDL 8.2.2 released By: tom.grydeland on Thu, 21 March 2013 02:38
		Re: IDL 8.2.2 released By: Mark Piper on Thu, 21 March 2013 09:41
		Re: IDL 8.2.2 released By: chris_torrence@NOSPAM on Mon, 15 April 2013 21:08
		Re: IDL 8.2.2 released By: tom.grydeland on Fri, 26 April 2013 03:06
		Re: IDL 8.2.2 released By: chris_torrence@NOSPAM on Fri, 26 April 2013 14:33
		Re: IDL 8.2.2 released By: tushar on Thu, 25 April 2013 23:31

Previous Topic:	netCDF adding variable to an existing file
Next Topic:	nearest node of Delauny tesselation

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Wed Oct 08 18:01:05 PDT 2025

Total time taken to generate the page: 0.00563 seconds