使用 OpenMP(Thread-Level Parallelism)，来操作矩阵和，点积
OpenMP stands for Open specification for Multi-Processing
Summary of OpenMP 3.0 C/C++ Syntax
#pragma omp parallel for 指令用于并行化 for 循环，而 reduction(+:total_sum) 部分告诉 OpenMP 在所有线程之间对 total_sum 变量执行加法归约操作。这意味着每个线程都会计算 total_sum 的一个局部值，然后 OpenMP 会在所有线程中归约（合并）这些局部值，最终得到全局的 total_sum。
Naive: 1 thread took 4.577141 seconds
Manual Optimized: 1 thread(s) took 0.875618 seconds
Manual Optimized: 2 thread(s) took 0.445462 seconds
Manual Optimized: 3 thread(s) took 0.314087 seconds
Manual Optimized: 4 thread(s) took 0.296813 seconds
Manual Optimized: 5 thread(s) took 0.320832 seconds
Manual Optimized: 6 thread(s) took 0.294578 seconds
Manual Optimized: 7 thread(s) took 0.287689 seconds
Manual Optimized: 8 thread(s) took 0.289076 seconds
Reduction Optimized: 1 thread(s) took 0.877703 seconds
Reduction Optimized: 2 thread(s) took 0.447750 seconds
Reduction Optimized: 3 thread(s) took 0.314114 seconds
Reduction Optimized: 4 thread(s) took 0.299070 seconds
Reduction Optimized: 5 thread(s) took 0.305538 seconds
Reduction Optimized: 6 thread(s) took 0.297613 seconds
Reduction Optimized: 7 thread(s) took 0.289809 seconds
Reduction Optimized: 8 thread(s) took 0.291706 seconds
Congrats! All dotp tests passed