weight_utils.cc File Reference

#include <vector>
#include <numeric>
#include <algorithm>
#include "healpix_map.h"
#include "alm_healpix_tools.h"
#include "weight_utils.h"
#include "alm.h"
#include "libsharp/sharp_cxx.h"
#include "lsconstants.h"

Go to the source code of this file.

Functions
vector< double >	get_fullweights (int nside, int lmax, double epsilon, int itmax, double &epsilon_out)
vector< double >	get_ringweights (int nside, int lmax, double epsilon, int itmax, double &epsilon_out)

Detailed Description

Functionality for computing ring weights and full map weights

Helpful literature: Graef, Kunis, Potts: On the computation of nonnegative quadrature weights on the sphere (https://www-user.tu-chemnitz.de/~potts/paper/quadgewS2.pdf) Lambers: Minimum Norm Solutions of Underdetermined Systems (http://www.math.usm.edu/lambers/mat419/lecture15.pdf) Shewchuk: An Introduction to the Conjugate Gradient Method Without the Agonizing Pain (https://www.cs.cmu.edu/~quake-papers/painless-conjugate-gradient.pdf)

Copyright (C) 2016-2019 Max-Planck-Society

Author

Martin Reinecke

Definition in file weight_utils.cc.

Function Documentation

get_fullweights()

vector<double> get_fullweights	(	int	nside,
		int	lmax,
		double	epsilon,
		int	itmax,
		double &	epsilon_out
	)

Computes full weights for a map of the given nside at a maximum multipole lmax. The solution is obtained via CGNE iteration, which stops once the norm of the residual falls below epsilon times the norm of the initial residual, or once itmax iterations have been performed.

Returns

a vector containing the compressed full weights.

Note

lmax must be even; odd l do not contribute to the weights.

Definition at line 340 of file weight_utils.cc.

get_ringweights()

vector<double> get_ringweights	(	int	nside,
		int	lmax,
		double	epsilon,
		int	itmax,
		double &	epsilon_out
	)

Computes ring weights for a map of the given nside at a maximum multipole lmax. The solution is obtained via CGNE iteration, which stops once the norm of the residual falls below epsilon times the norm of the initial residual, or once itmax iterations have been performed.

Returns

a vector of size 2*nside containing the ring weights.

Note

lmax must be even; odd l do not contribute to the weights.

Definition at line 380 of file weight_utils.cc.