Spline Contour - Spline Contour - Сообщения
#1 Опубликовано: 12.09.2016 14:36:46
This one works well for the given data set. Mesh finer at will and if needed,
bilinear ... not much use for bilinear, whereas "spline" is just something.
Jean
Spline Esplina [AI2000].sm
bilinear ... not much use for bilinear, whereas "spline" is just something.
Jean
Spline Esplina [AI2000].sm
#2 Опубликовано: 12.09.2016 19:08:01
... a faster version: one twin col extractor removed.
By same token, I wonder why the col extractor col(,,) is so slow ?
Jean
Spline Esplina [AI2000].sm
By same token, I wonder why the col extractor col(,,) is so slow ?
Jean
Spline Esplina [AI2000].sm
#3 Опубликовано: 12.09.2016 21:59:52
"By same token, I wonder why the col extractor col(,,) is so slow ?"
You can speed up calculation if o are using "eval":
Spline Esplina [AI2000] (1).sm
You can speed up calculation if o are using "eval":
Spline Esplina [AI2000] (1).sm
#4 Опубликовано: 12.09.2016 23:07:59
Thanks Ber7 for "speedy Gonzalez". In fact, the "uN"
conditional code is redundant in this dataset.
Here is the latest version, even faster.
Jean
Spline Esplina [AI2000].sm
conditional code is redundant in this dataset.
Here is the latest version, even faster.
Jean
Spline Esplina [AI2000].sm
#5 Опубликовано: 13.09.2016 03:08:10
Jean,calculation speed will be doubled, if we use the advice of Андрей
http://en.smath.info/forum/resource.ashx?a=14504&b=2
Spline Esplina [AI2000] 2.sm
http://en.smath.info/forum/resource.ashx?a=14504&b=2
Spline Esplina [AI2000] 2.sm
1 пользователям понравился этот пост
Davide Carpi 13.09.2016 11:45:00
#7 Опубликовано: 13.09.2016 11:23:31
You forgot to specify first element of u vector.
By doing so you can speed up calculation even more:
Regards.
By doing so you can speed up calculation even more:
Regards.
2 пользователям понравился этот пост
#8 Опубликовано: 13.09.2016 12:08:37
You forgot to specify first element of u vector.
By doing so you can speed up calculation even more:
Thanks Andrey, damned me !
While you are there, can you speed up "DeBoor" ?
Jean
Spline De Boor [Esplina].sm
#9 Опубликовано: 13.09.2016 12:47:49
Done
Spline De Boor [Esplina]_mod.sm
Timings:
Regards.
Spline De Boor [Esplina]_mod.sm
Timings:
- 6.1 seconds by 0.97.5890
- 2.3 seconds by 0.98.6096
Regards.
#10 Опубликовано: 13.09.2016 15:32:56
Thanks Andrey,
... for the local definition of Sx(v), Sy(v).
Must be consistant with all upward local definitions.
Jean
... for the local definition of Sx(v), Sy(v).
Must be consistant with all upward local definitions.
Jean
#11 Опубликовано: 13.09.2016 15:53:06
You're welcome.
The main change was this:
Actually I'm not sure why calculation of u is so complex... looks like it is just a copy of m.
The main change was this:
Actually I'm not sure why calculation of u is so complex... looks like it is just a copy of m.
#12 Опубликовано: 14.09.2016 00:51:51
Quite right Andrey: 'u' is 'm', which 'm' is redundant.
I must have constructed N_(v,j) aroud 'u' from some source,
thus easier to check the coding. The attached work sheet
has been refactored with auxilliary explanations.
Can be used as model for DeBoor, pretty easy to adapt to
any project. Spline order is linear interpolation around
the polygon. Spline order 2 is "convex hull" per say.
They are about the only two types splines useful for
technical works [?]
Jean
Spline De Boor [Loop].sm
I must have constructed N_(v,j) aroud 'u' from some source,
thus easier to check the coding. The attached work sheet
has been refactored with auxilliary explanations.
Can be used as model for DeBoor, pretty easy to adapt to
any project. Spline order is linear interpolation around
the polygon. Spline order 2 is "convex hull" per say.
They are about the only two types splines useful for
technical works [?]
Jean
Spline De Boor [Loop].sm
- Новые сообщения
- Нет новых сообщений