Floating Point Unit Verilog

IEEE 754 Floating Point Unit written in Verilog

Requirements:

• g++ compiler
• Modelsim

How to run (Windows)

1. Compile test.cpp file using g++ compiler.

g++ -o test.exe test.cpp

For more information about the program after compilation, type

.\test.exe --help

2. Generate test case using:

.\test.exe --gen

• On your terminal you should see:

Number of cases to generate:

• Enter the number of cases you want to generate. Let's take an example of 10 The output should be:

Number of cases to generate:  10
Generating...
Input testbench file generated successfully!
Input testbench file are saved as tb_input_gen.txt

• You should see file tb_input_gen.txt and tb_output_gen.txt in your directory.

3. Simulation:

• Open Modelsim
• Change project directory to this code folder
• Add all the verilog code to the project
• In file fpu_tb.v on line 2, you should see:

parameter NUM_CASES = 10;

• Change the NUM_CASES according to the number of cases you gen before.
• Compile the code and simualte.
• (Optional) If you want to see waveform of the FPU, you go to File -> Load -> Macro File and choose file wave.do from the project directory.
• Use run -all command in Verilog console or press Run -All button in the toolbar.
• The code will automatically stop running after loop through all test cases.
• You should see an output file named output_tb_fpu.txt in your directory.

4. Evaluation:

• Because of our Verilog is not good enough, you have to manually delete the first line of FPU generated file named output_tb_fpu.txt
• Rerun the test.exe program with the following command:

.\test.exe --eval

• The terminal will display the nunber of errors if any along with the line that is differences from ground truth.
• If all the test cases are passed you will see Test passed! on your terminal.