examine points which are inside a polygon by GPU

Hi All,

Does anyone know how I can speed up inpolygon function by using GPU?

Thanks.

Answer 1

[Sulimon Sattari]

I wrote an inpolygon GPU function on matlabcentral.

This will only give you speed-up if the number of test points is much much greater than the number of vertices (say, 10,000 times more test points than vertices).

http://www.mathworks.com/matlabcentral/fileexchange/49157-inpolygon-for-gpu

Regards, Sulimon

Answer 2

[Walter Roberson]

I thought someone asked about this about a week ago, but I cannot find the posting now. The answer was that TriScatteredInterp's routine for checking points in polygons was not designed for GPU use.

Some faster point-in-polygon functions:

http://www.mathworks.com/products/parallel-computing/demos.html?file=/products/demos/distribtb/MapDemo/MapDemo.html

http://www.mathworks.com/matlabcentral/fileexchange/10391-fast-points-in-polygon-test

If, though, as the other person did, you specifically want to test whether a particular point is outside all of the polygons created by TriScatteredInterp, then you would probably have to write the GPU version of the routine yourself.

[AP]

I have seen these links. In the first link, the function does not exist in my MATLAB. Is it the case in the version of your MATLAB? Can you call the function like the following:

[in,t] = inpoly(xGrid,yGrid,xv,yv,useGPU)

I am also trying to convert the code in the link http://www.mathworks.com/matlabcentral/fileexchange/10391-fast-points-in-polygon-test to GPU version. Could you please give me some hints?

[Walter Roberson]

As the first demo is about mapping, probably inpoly() is part of the mapping toolbox.

I have no experience in writing GPU ccode.