function [ x, error_norm, iter, flag ] = cg ( A, x, b, M, max_it, tol ) %% CG solves a linear system using the conjugate gradient method. % % Discussion: % % The algorithm is appropriate for a symmetric positive definite matrix A. % % This routine does NOT carry out preconditioning. % % Modified: % % 19 July 2004 % % Reference: % % Richard Barrett, Michael Berry, Tony Chan, James Demmel, % June Donato, Jack Dongarra, Victor Eijkhout, Roidan Pozo, % Charles Romine, Henk van der Vorst % Templates for the Solution of Linear Systems: Building Blocks for % Iterative Methods, % SIAM Publications, 1993. % % Parameters: % % Input, real A(N,N), the symmetric positive definite matrix. % % Input, real X(N), the initial guess vector. % % Input, real B(N), the right hand side vector. % % Input, real M, not used. % % Input, integer MAX_IT, the maximum number of iterations. % % Input, real TOL, an error tolerance. % % Output, real X(N), the solution. % % Output, real ERROR_NORM, the norm of the error. % % Output, integer ITER, the number of iterations performed. % % Output, integer FLAG, the return flag. % 0 = the solution was found to within the specified tolerance. % 1 = a satisfactory solution was not found. The iteration limit % was exceeded. % flag = 0; iter = 0; bnrm2 = norm ( b ); if ( bnrm2 == 0.0 ) bnrm2 = 1.0; end r = b - A * x; error_norm = norm ( r ) / bnrm2; errorhist = [ ]; errorhist(1) = error_norm; if ( error_norm < tol ) return end for iter = 1 : max_it z = r; rho = ( r' * z ); % % Compute P, the direction vector. % if ( 1 < iter ) beta = rho / rho_1; p = z + beta * p; else p = z; end q = A * p; alpha = rho / ( p' * q ); % % Update the approximate solution. % x = x + alpha * p; % % Compute the residual. % r = r - alpha * q; error_norm = norm ( r ) / bnrm2; errorhist(iter+1) = error_norm; if ( error_norm <= tol ) break end rho_1 = rho; end if ( tol < error_norm ) flag = 1; end error_norm = errorhist; return end