Digital Arithmetic By Ercegovac And Lang | Pdf

Let’s use a known correct mapping: Decimal 7 in 4-bit binary: 0111. SD: 1001 (1×8 + (-1)×4 + 0×2 + 1×1) = 8 – 4 + 1 = 5. No.

Let’s simplify: A correct SD radix-2 example: Decimal 5: binary 0101. SD: 1101? 1×8 + (-1)×4 + 1×2 + 1×1 = 8 – 4 + 2 + 1 = 7. Still 7. digital arithmetic by ercegovac and lang pdf

Below is an original feature titled: —inspired by themes from Ercegovac & Lang (e.g., redundant number systems, signed-digit representations, and online arithmetic). Feature: Recoding and Redundancy – The Secret to High-Speed Arithmetic 1. The Problem with Conventional Addition In standard binary addition, carry propagation limits speed. Adding two n -bit numbers in worst case requires O( n ) gate delays due to the ripple carry. Even carry-lookahead adders face practical limits as n grows. Let’s use a known correct mapping: Decimal 7

If we allow digits to be redundant (e.g., digit set {-1,0,1} instead of {0,1}), addition becomes carry-free within a small constant window. 2. Introducing Redundant Signed-Digit (SD) Representation A radix- r signed-digit number uses digit set { -α, …, α } where α > r/2. For radix 2, the digit set {-1,0,1} works. Let’s simplify: A correct SD radix-2 example: Decimal

Better example: Decimal 3 in binary: 0011 (3). SD representation: 0101? 0×4 + 1×2 + (-1)×1? That’s 1. Not right.

The decimal number 5 in 4-bit binary is 0101 . In SD (radix-2, digits -1,0,1), 5 can be represented as 0101 (same) or 1011 (where 1 means -1 at that position). Let’s verify: 1011 (SD) = 1×8 + (-1)×4 + 1×2 + 1×1 = 8 – 4 + 2 + 1 = 7? Wait, that’s 7, not 5 — so not correct. Let’s do properly: