I developed the program step by step, function by function, and eventually combined them into the final result. Below is the usage of each function listed in the order in which I created them:

1. convolution.m - MATLAB code to simulate the result of convolution and display the results in a chart.
2. convolution_origin.c - The original convolution code for floating-point numbers.
3. fp16tfp32.c - For understanding the transformation from FP16 to FP32.
4. fp32tfp16.c - For understanding the transformation from FP32 to FP16.
5. fp32_mul.c - Referenced from the Quiz 1 of last year.
6. fp16_mul.c - Adapted from fp32_mul.c to handle FP16.
7. convolution_fp16_v1.c - Before successfully implementing fp16_add, I converted the multiplication result to FP32 and then added them together.
8. fp16_add.c - Handcrafted bitwise addition for FP16 convolution.
9. convolution_fp16_v2.c - Combined all the functions to produce the final result for FP16 convolution.