# Accumulator Error Feedback

### From Wikimization

(Difference between revisions)

Line 19: | Line 19: | ||

% % Matlab csum() example: | % % Matlab csum() example: | ||

% clear all | % clear all | ||

- | % csumv = 0; | + | % csumv=0; rsumv=0; |

- | % while | + | % while csumv <= rsumv |

% v = randn(13e6,1); | % v = randn(13e6,1); | ||

% rsumv = abs(sum(v) - sum(v(end:-1:1))); | % rsumv = abs(sum(v) - sum(v(end:-1:1))); | ||

Line 39: | Line 39: | ||

return | return | ||

</pre> | </pre> | ||

+ | |||

+ | === notes === | ||

+ | In practice, input sorting | ||

+ | <pre> | ||

+ | [~, idx] = sort(abs(x),'descend'); | ||

+ | x = x(idx); | ||

+ | </pre> | ||

+ | should begin the <tt>csum()</tt> subroutine to achieve a most accurate summation. | ||

+ | That is not presented here because the commented example (inspired by Higham) would then display false positive results. | ||

+ | Even in absence of sorting, <tt>csum()</tt> is more accurate than conventional summation. | ||

=== links === | === links === |

## Revision as of 23:01, 28 November 2017

function s_hat = csum(x) % CSUM Sum of elements using a compensated summation algorithm. % % For large vectors, the native sum command in Matlab does % not appear to use a compensated summation algorithm which % can cause significant roundoff errors. % % This code implements a variant of Kahan's compensated % summation algorithm which often takes about twice as long, % but produces more accurate sums when the number of % elements is large. -David Gleich % % Also see SUM. % % % Matlab csum() example: % clear all % csumv=0; rsumv=0; % while csumv <= rsumv % v = randn(13e6,1); % rsumv = abs(sum(v) - sum(v(end:-1:1))); % disp(['rsumv = ' num2str(rsumv,'%18.16f')]); % [~, idx] = sort(abs(v),'descend'); % x = v(idx); % csumv = abs(csum(x) - csum(x(end:-1:1))); % disp(['csumv = ' num2str(csumv,'%18.16e')]); % 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 = (s_hat_old - s_hat) + y; %calculate difference first (Higham) end return

### notes

In practice, input sorting

[~, idx] = sort(abs(x),'descend'); x = x(idx);

should begin the `csum()` subroutine to achieve a most accurate summation.
That is not presented here because the commented example (inspired by Higham) would then display false positive results.
Even in absence of sorting, `csum()` is more accurate than conventional summation.

### links

Accuracy and Stability of Numerical Algorithms 2e, ch.4.3, Nicholas J. Higham, 2002

For multiplier error feedback, see:

Implementation of Recursive Digital Filters for High-Fidelity Audio

Comments on Implementation of Recursive Digital Filters for High-Fidelity Audio