Floating Point Multiplication and Division

using RISC-V Assembly

Method

Write RISC-V321 base integer assembly instruction code to implement a function for floating-point multiplication. Similarly, also write assembly code to implement a function for floating-point division. Use IEEE 754 format single precision floating point representation.

You will debug and simulate your code using a simulator such as Venus, https://

www.kvakil.me/venus/ . Details on using Venus are given at https://github.com/kvakil/venus/

wiki .

Project Description

You will create a main routine which calls 2 floating point subroutines: fmult and fdiv. You will

submit a .txt file that includes all the code

demonstrate your code using the following values:

multiplicand: 8.0546875 100

multiplier: -1.79931640625 10-1

dividend: 8.625 101

division: -4.875 100

Part 1: fmult

Develop and test RISC-V assembly code to implement floating-point multiplication. Use the

floating-point multiplication algorithm given in Figure 3.16 of the reference textbook (on reserve

at the Engineering Library) as the basis for your floating point multiplication routine. Return a

32-bit product. Use the simulator for debugging and simulation. Use round to nearest even

with 1 guard bit, 1 round bit, and 1 sticky bit. You should also make sure that you normalize

properly. You do not need to handle denormals or special values other than zero.

Part 2: fdiv

Develop and test RISC-V assembly code to implement floating-point division. Use an algorithm

similar to the multiplication algorithm you use in Part 1. Return a 32-bit quotient and a 32-bit remainder. Again, use the simulator for debugging and simulation. Use round to nearest even

with 1 guard bit, 1 round bit, and 1 sticky bit. You should also make sure that you normalize

properly. You do not need to handle denormals or special values other than zero.

Part 3: fmult and fdiv integrated called from main

Write a main program using RISC-V assembly that (1) reads the test values from memory, (2)

calls fmult and stores the result in memory, and (3) calls fdiv and stores the result in memory.

You should convert the multiplicand, multiplier, dividend and divisor to IEEE single-precision

floating point format by hand and store it in a suitable memory location. Your resulting product

should be stored in memory using single-precision floating point.