Download presentation
Presentation is loading. Please wait.
Published byEzra Gordon Modified over 8 years ago
1
TK2633 : MICROPROCESSOR & INTERFACING Lecture 10: Fixed Point Arithmetic Lecturer: Ass. Prof. Dr. Masri Ayob
2
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob2 Fixed Point Arithmetic. 8-bit binary addition. Suppose that the ten 8-bit numbers, stored at memory locations 2800H through 2809H, are added. To sum these data, an 8-bit instructions are chosen. The ADD M instruction is chosen. Since this is to add 10 numbers, the programme loop construct is ideal for this problem. Figure illustrates the flowchart to tackle the problem.
3
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob3 The 8-bit binary addition listed in the example below, results in a problem that the largest sum can be only 0FFH. This sum is appropriate only if 10 numbers are small, therefore to avoid overflow for bigger sum operation, the 16-bit sum operation will take place. 8-bit binary addition
4
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob4 Example of 16-bit summation. 8-bit binary addition
5
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob5 Suppose that two lists of numbers, each 10H bytes long, appear in the memory to be subtracted. A programme must take the number stored at LIST2 and subtract it from the number stored at LIST1. The difference must be stored at a location in LIST2. This operation repeats 10H times until all sets of numbers are subtracted. Figure illustrates the process. 8-bit binary subtraction
6
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob6 8-bit binary subtraction: Exercise There are two lists of numbers (LIST1 and LIST2), each 10H bytes long, appear in the memory. Write a program to take the number stored at LIST2 and subtract it from the number stored at LIST1. The difference must be stored at a location in LIST2.
7
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob7 Repeated addition is the simplest method of multiplication to understand. Suppose to multiply two 8- bit numbers, one number can be used as the loop counter. The other number can be added the number of times stored in the loop counter. Figure illustrates how the number 6 can be multiplied by 3. Unsigned Multiplication by Repeated Addition
8
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob8 Unsigned Multiplication by Repeated Addition
9
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob9 Unsigned Constant Multiplication Instead of using the repeated addition technique presented earlier, another technique often results in a much faster multiplication, which using DAD instruction. The easiest way to multiply by 4 is by using the DAD H instruction. Remember that if double HL twice, it is multiplied by 4. Example :
10
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob10 The most flexible version of multiplication is the unsigned multiplication shift and add algorithm. Observe the example below, the unsigned multiplication is a combination of shifting the number followed by addition for every shifted number. Figure below, illustrates the process of shifting and adding number 9 and 15. Unsigned Multiplication Algorithm
11
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob11 The flowchart example to perform the algorithm. Unsigned Multiplication Algorithm
12
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob12 Unsigned Multiplication Algorithm
13
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob13
14
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob14 Modified Flowchart diagram to perform signed multiplication. Signed Multiplication
15
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob15 To do the signed multiplication, the multiplicand and multiplier are exclusive ORed together. The result, which indicates the sign of the product, is saved in the B register. Both numbers are then adjusted so that they are positive before the MULT subroutine is called. Upon returning from the MULT subroutine, the sign of the B register shows the sign of the product. For a negative result, the product in the HL register pair must be negated to form a negative product. For a positive result, the product does not change. Signed Multiplication
16
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob16 Signed Multiplication: Modification of the previous program
17
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob17 If data are shifted to the left, they multiply by 2 for each bit position of the shift. If data are shifted to the right, they divide by 2 for each bit position of the shift. Knowing this as a technique used to multiply by a constant, a technique can be developed so that a number can be divided by any power of 2. Division by a constant
18
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob18
19
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob19 Example shows how the signed contents of the accumulator are divide by 4 with the result rounded. Notice that a shift left copies the sign bit into the carry flag before shifting right twice. Signed division by a constant
20
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob20 To divide a number by any integer value, a division algorithm is normally used to develop a subroutine. The division algorithm uses a combination of shifting left, comparing, subtracting, and setting bits to perform binary division. Below is an illustration of division algorithm. Unsigned division algorithm
21
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob21 Unsigned division algorithm
22
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob22 Unsigned division algorithm
23
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob23 Signed binary division is treated in the same manner as signed multiplication. Signed division algorithm
24
June 4, 2016 Written by: Ramizi Mohamed, Edited by:Dr Masri Ayob24 Thank you Q&A
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.