<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<item>
  <id>05336911</id>
  <dt>j</dt>
  <an>05336911</an>
  <augroup>
    <au>Watanuki, O.</au>
    <au>Ercegovac, M.D.</au>
  </augroup>
  <ti>Error analysis of certain floating-point on-line algorithms.</ti>
  <so>IEEE Trans. Comput. 32, No. 04, 352-358 (1983).</so>
  <py>1983</py>
  <pu>Institute of Electrical and Electronics Engineers (IEEE), Washington, DC</pu>
  <lagroup>
    <la>EN</la>
  </lagroup>
  <ccgroup>
  </ccgroup>
  <utgroup>
    <ut>redundant significand</ut>
    <ut>absolute error</ut>
    <ut>floating-point on-line algorithm</ut>
    <ut>maximum relative representation error</ut>
    <ut>pseudo-normalized significand</ut>
  </utgroup>
  <cigroup>
  </cigroup>
  <ligroup>
    <li>doi:10.1109/TC.1983.1676236</li>
  </ligroup>
  <abgroup>
    <ab>Summary: The properties of redundant number system in significand (mantissa) representation are studied and the range of redundant significand is derived. From the range of the redundant significand and the absolute error of on-line operations, the MRRE (maximum relative representation error) is defined and analyzed for floating-point on-line addition and multiplication.</ab>
    <rv></rv>
  </abgroup>
</item>