Sample Reference Markup
 Home 

Here (below the horizontal rule) is the markup used for the sample reference (with a blank on the start of each line which prevents it being formatted here).
 <nick>cowlis2003
 <type>paper
 <cat>fboqsv

 <title>Decimal Floating-Point: Algorism for Computers
 <author>Cowlishaw, Michael F.
 <in>Proceedings of the 16th IEEE Symposium on Computer Arithmetic
 <isbn>0-7695-1894-X
 <pub>IEEE
 <pp>104-111
 <date>June 2003
 <URL>http://www.dec.usc.es/arith16/papers/paper-107.pdf
 <file>cowlis2003-DFP-algorism.pdf

 <has>
 Decimal floating-point justification and new combined arithmetic rules.

 <abstract>
 Decimal arithmetic is the norm in human calculations, and
 human-centric applications must use a decimal floating-point
 arithmetic to achieve the same results.

 Initial benchmarks indicate that some applications spend 50% to 90% of
 their time in decimal processing, because software decimal arithmetic
 suffers a 100&#215; to 1000&#215; performance penalty over hardware.
 The need for decimal floating-point in hardware is urgent.

 Existing designs, however, either fail to conform to modern standards
 or are incompatible with the established rules of decimal arithmetic.
 This paper introduces a new approach to decimal floating-point which
 not only provides the strict results which are necessary for
 commercial applications but also meets the constraints and
 requirements of the IEEE 854 standard.

 A hardware implementation of this arithmetic is in development, and it
 is expected that this will significantly accelerate a wide variety of
 applications.

 <note>
 Softcopy is available in
 <a href="http://www.dec.usc.es/arith16/papers/paper-107.pdf">PDF</a>.


Sample Reference Markup was last edited on 2006-04-24  
Project: MemoWiki