Prev: Refresh Simulink library browser
Next: Using Apache CXF in from MATLAB
From: Paul on 28 Jul 2010 22:33
"Paul " <paulmark(a)eden.rutgers.edu> wrote in message <i2qns3$7l8$1(a)fred.mathworks.com>...
> Corrections (again):
>
> @Bruno - sorry, i missed the mention of interp1q in the link you sent me. A 20x speedup would be wonderful! Going to investigate that tomorrow.

Just compiled the mex file on a 32bit system. >3x speedup is very impressive. Thank you very much!
From: Bruno Luong on 29 Jul 2010 01:30
"Paul " <paulmark(a)eden.rutgers.edu> wrote in message <i2qngr$f1n$1(a)fred.mathworks.com>...

>
> @Bruno - Ill look at that mex file tomorrow. What are the inputs it takes? Same as interp1? Is it like interp1q?

The input are *column* vectors, (x,y) are coarse data, with x sorted in ascending order, and xi is the abscissa where interpolation takes place also in column. The MEX does not verify those conditions to avoid overhead so enhance the speed.

>I don't know what you are talking about with HISTC being the default. For my Matlab it is the cumulative histogram and im honestly not clever enough to get from histograms to searching sorted lists.

The second output of HISTC (loc) tells where the data is located in the sorted edges.

Bruno
First  |  Prev  | 
Pages: 1 2 3
Prev: Refresh Simulink library browser
Next: Using Apache CXF in from MATLAB