and eventually scalar operations. Hopefully this first set of comparisons demonstrates the power of optimization flags, and what happens when a compiler is able to auto-vectorize your code. Even ignoring the unrealistic absence of any optimization, we were able to see a roughly four
4x Code Performance with SIMD