Report Number: CS-TR-70-163
Institution: Stanford University, Department of Computer Science
Title: An algorithm for floating-point accumulation of sums with small relative error
Author: Malcolm, Michael A.
Date: June 1970
Abstract: A practical algorithm for floating-point accumulation is presented. Through the use of multiple accumulators, errors due to cancellation are avoided. An example in Fortran is included. An error analysis providing a sharp bound on the relative error is also given.
http://i.stanford.edu/pub/cstr/reports/cs/tr/70/163/CS-TR-70-163.pdf