Very slow Smoothing function

TeeChart VCL for Borland/CodeGear/Embarcadero RAD Studio, Delphi and C++ Builder.
Post Reply
moelski
Advanced
Posts: 212
Joined: Mon Apr 23, 2007 12:00 am
Location: Germany
Contact:

Very slow Smoothing function

Post by moelski » Fri May 04, 2007 11:49 am

Hi !

I did some tests with the smoothing function because we need it for curve smoothing.

What I tested was the following:
- new application
- add TeeChart component
- add 2 FastLine Series to the Chart
- Series -> Data Source
> set Series1 to Random (1000 values)
> set Series2 to Function -> Smoothing (Series: Series1)
> click Apply
If you click the apply button as the last step, TeeChart takes a really long time for calculation. The problem gets bigger if you increase the random values to 10000. The application will hang for a long time.

Is this a bug? I can´t imagine why it takes such a long time to calculate a smooth (interpolated) line.

Any hints?

Pep
Site Admin
Site Admin
Posts: 3299
Joined: Fri Nov 14, 2003 5:00 am
Contact:

Post by Pep » Mon May 07, 2007 4:47 pm

Hi Greetz,

it's not a bug, the code used to calculate the spline takes huge CPU time with more than 500 points.
I've added a note on out wish list ( try to improve the speed ) to be considered for further releases.

moelski
Advanced
Posts: 212
Joined: Mon Apr 23, 2007 12:00 am
Location: Germany
Contact:

Post by moelski » Tue May 08, 2007 5:05 am

Hi Pep,

thx for that info.

When do you think is this topic fixed?

Greetz Dominik

btw: May name isn´t Greetz :wink: Greetz = Greetings

Pep
Site Admin
Site Admin
Posts: 3299
Joined: Fri Nov 14, 2003 5:00 am
Contact:

Post by Pep » Tue May 08, 2007 11:09 am

Hi Dominik,

I'm sorry, I annot tell for sure when will be check it in order to be improved. If you are "source code" customer you can take a look the TeeSpline unit and try to find a way to speed it.

>>> btw: May name isn´t Greetz Greetz = Greetings
Ups !.. sorry :wink:

moelski
Advanced
Posts: 212
Joined: Mon Apr 23, 2007 12:00 am
Location: Germany
Contact:

Post by moelski » Tue May 08, 2007 12:47 pm

Hi Pep,

Ok I will take a look. We have the source code.

But I think we will mplement our own calculation.
Anyway ... It would be great if you could fix this problem in the future.

Greetz

Post Reply