Unformatted text preview: ram performance. 3.14.2 ExtendedPrecision Arithmetic
A second unusual feature of IA32 ﬂoating point is that the ﬂoatingpoint registers are all 80 bits wide. They encode numbers in an extendedprecision format as described in Problem 2.49. It is similar to an IEEE ﬂoatingpoint format with a 15bit exponent (i.e., ½ ) and a 63bit fraction (i.e., Ò ¿). All single and doubleprecision numbers are converted to this format as they are loaded from memory into ﬂoatingpoint registers. The arithmetic is always performed in extended precision. Numbers are converted from extended precision to single or doubleprecision format as they are stored in memory. This extension to 80 bits for all register data and then contraction to a smaller format for all memory data has some undesirable consequences for programmers. It means that storing a value in memory and then retrieving it can change its value, due to rounding, underﬂow, or overﬂow. This storing and retrieving is not always visible to the C programmer, lea...
View
Full
Document
This note was uploaded on 09/02/2010 for the course ELECTRICAL 360 taught by Professor Schultz during the Spring '10 term at BYU.
 Spring '10
 Schultz
 The American, Gulliver's Travels, 2.2.5 2.2.6 2.2.7 2.3 2.3.1 2.3.2 2.3.3 2.3.4 2.3.5

Click to edit the document details