here is a python library for
polynomial manipulation presented in the section
(
Piecewise polynomials
). In this
section we optimize for speed using Cuda.
The source code may be downloaded from the Download section. It includes make
files and MS Visual Studio project files. The make files build a release
extension for Python with full functionality. The MSVS project files build a
debug executable file with some testing functionality.
We restrict our attention to needs of wavelet analysis, see the section
(
Wavelet analysis section
). In
particular, the presented here library performs well for piecewise polynomials
of small degree and large number of pieces and it does not perform too well
for large degree and small number of pieces.
There is a __shared__ memory-optimized code inside
-function
of ots::poly::impl namespace that refuses to work if the degree is greater
than or equal to 16. This limitation may be removed by making localized
corrections within that function.
We characterize a piecewise polynomial
,
with the following set of
parameters:
|
|
(Piecewise polynomial representation)
|
See the section (
Wavelet analysis
section
) for the notation
.
|
