PC20101110

# PC20101110 - Numerical Analysis II Dr Abigail Wacher Nov...

This preview shows pages 1–2. Sign up to view the full content.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Numerical Analysis II Dr Abigail Wacher Nov 10, 2010 1 Let us use a decimal floating point system F with 4 digits and an exponent between -9 and 9. b = 10, t = 4, u = K L = 9 (Associativity) Take a = 1.000 \$ 10 , b = 3.000 \$ 10 K 4 , c = 4.000 \$ 10 K 4 a C b = 1.000 \$ 10 C 0.0003 \$ 10 = 1.000 \$ 10 a C b C c = 1.000 \$ 10 C 0.0004 \$ 10 = 1.000 \$ 10 b C c = 3.000 \$ 10 K 4 C 4.000 \$ 10 K 4 = 7.000 \$ 10 K 4 a C b C c = 1.000 \$ 10 C 0.0007 \$ 10 = 1.001 \$ 10 Not associative! Uniqueness of solutions to algebraic equations does not hold for floating point arithmetic. 1 C x = 1 has more than 1 solution in F not just the "exact" solution x = 0. In fact in F , every positive floating point number y ! 5.000 \$ 10 K 4 is a correct solution to the equation. In the course of solving ax 2 K 2 bx C c = 0, for x , the expression b 2 K ac must be computed. Can the true value of b 2 K ac be non-negative and yet have a negative computed value?...
View Full Document

## This note was uploaded on 02/09/2011 for the course MATH 2051 taught by Professor Dr.a.wacher during the Fall '11 term at Durham.

### Page1 / 2

PC20101110 - Numerical Analysis II Dr Abigail Wacher Nov...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online