Accumulator Error Feedback
From Wikimization
(→sorting) |
|||
Line 55: | Line 55: | ||
In practice, input sorting can sometimes achieve more accurate summation. | In practice, input sorting can sometimes achieve more accurate summation. | ||
Sorting later became integral to modifications of Kahan's algorithm, such as Priest's | Sorting later became integral to modifications of Kahan's algorithm, such as Priest's | ||
- | (Higham Algorithm 4.3). | + | ([http://servidor.demec.ufpr.br/CFD/bibliografia/Higham_2002_Accuracy%20and%20Stability%20of%20Numerical%20Algorithms.pdf Higham] Algorithm 4.3). |
But the very same accuracy dependence on input data prevails. | But the very same accuracy dependence on input data prevails. | ||
Revision as of 22:16, 22 February 2018
function s_hat = csum(x) % CSUM Sum of elements using a compensated summation algorithm. % % This Matlab code implements % Kahan's compensated summation algorithm (1964) % which takes about twice as long as sum() but % produces more accurate sums when number of elements is large. % -David Gleich % Also see SUM. % % Example: % clear all; clc % csumv=0; rsumv=0; % n = 100e6; % t = ones(n,1); % while csumv <= rsumv % v = randn(n,1); % % rsumv = abs((t'*v - t'*v(end:-1:1))/sum(v)); % disp(['rsumv = ' num2str(rsumv,'%1.16f')]); % % csumv = abs((csum(v) - csum(v(end:-1:1)))/sum(v)); % disp(['csumv = ' num2str(csumv,'%1.16e')]); % end s_hat=0; e=0; for i=1:numel(x) s_hat_old = s_hat; y = x(i) + e; s_hat = s_hat_old + y; e = y - (s_hat - s_hat_old); end return
summing
ones(1,n)*v and sum(v) produce different results in Matlab 2017b with vectors having only a few hundred entries.
Matlab's variable precision arithmetic (VPA), (vpa(), sym()) from Mathworks' Symbolic Math Toolbox, can neither accurately sum a few hundred entries in quadruple precision. Error creeps up above |2e-16| for sequences with high condition number (heavy cancellation defined as large sum|x|/|sum x|). Use Advanpix Multiprecision Computing Toolbox for MATLAB, preferentially. Advanpix is hundreds of times faster than Matlab VPA. Higham measures speed here: https://nickhigham.wordpress.com/2017/08/31/how-fast-is-quadruple-precision-arithmetic
sorting
Floating-point compensated-summation accuracy is data dependent. Substituting a unit sinusoid at arbitrary frequency, instead of a random number sequence input, can make compensated summation fail to produce more accurate results than a simple sum.
In practice, input sorting can sometimes achieve more accurate summation. Sorting later became integral to modifications of Kahan's algorithm, such as Priest's (Higham Algorithm 4.3). But the very same accuracy dependence on input data prevails.
references
Accuracy and Stability of Numerical Algorithms 2e, ch.4.3, Nicholas J. Higham, 2002
Further Remarks on Reducing Truncation Errors, William Kahan, 1964
XSum() Matlab program - Fast Sum with Error Compensation, Jan Simon, 2014
For fixed-point multiplier error feedback, see:
Implementation of Recursive Digital Filters for High-Fidelity Audio
Comments on Implementation of Recursive Digital Filters for High-Fidelity Audio