Hi Soufiane, > My mistake is from a misunderstanding of the documentation. I read that: > > Conjugate Gradient (CG) symmetric positive definite > Stabilized Bi-CG (BiCGStab) non-symmetric > Generalized Minimum Residual (GMRES) general > > Like "General A" NxM, my is for any kind of square-matrices.
Ah, I see, I'll improve the wording. :-) > I have already try least square via QR, SVD... But the result are "too > sharp" (the result are visual position on 3D space), I would like a more > smooth method, that why I try to use a method based on gradient or > iterative method. > Did you know any method with a "relaxation-step" inside or based on > gradient I could use with rectangular matrices? As long as you are trying to solve the same least squares problem, you should get the same solution (otherwise the solver hasn't converged). To get a 'smoother' result, you can either apply weights to the individual points (search for 'weighted least squares') or use a general nonlinear distance metric, e.g. logarithmic. Best regards, Karli ------------------------------------------------------------------------------ Sponsored by Intel(R) XDK Develop, test and display web and hybrid apps with a single code base. Download it for free now! http://pubads.g.doubleclick.net/gampad/clk?id=111408631&iu=/4140/ostg.clktrk _______________________________________________ ViennaCL-devel mailing list ViennaCL-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/viennacl-devel
