[ncl-talk] Bug in "ezfftf" function in NCL version 6.1.2

[Mary Haley]

Dear Qian Li,

I'm sorry about the delay in response.  I've been out for a workshop and
then took an extended weekend.

I believe the error was with ezfftb and not ezfftf, and only in the case of
where you have an odd number of points. In V6.1.2 and earlier, ezfftb was
not using the "cf at npts" attribute correctly.  If "cf at npts" was even, then
it was okay. If it was odd, then the wrong size was used for some of the
internal work arrays, and hence the output results were incorrect.

I've attached a simple test program to help you see the issue. "xf" and
"xf_ans" should be equal to each other in V6.2.0 and later, but not with
V6.1.2 or earlier.

--Mary




On Fri, Jun 12, 2015 at 1:57 PM, Lee Evelyn <lee.evelyn.7 at gmail.com> wrote:

> Good afternoon,
>
> I have a question about the bug in function of "ezfftf". I used the
> earlier one in NCL version 6.1.2, but I saw the website said that there is
> a bug. I found that the output dimensions are different between those in
> version 6.1.2 (2N+1) and 6.3.0 (2N).  Can you tell me the specific errors
> in the earlier program? Because it may have a significant influence on my
> research results. Thanks a lot for your help.
>
> Best,
> Qian Li
>
> _______________________________________________
> ncl-talk mailing list
> ncl-talk at ucar.edu
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20150616/96168359/attachment.html 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: etest.ncl
Type: application/octet-stream
Size: 1242 bytes
Desc: not available
Url : http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20150616/96168359/attachment.obj