Is memory a bottleneck in matrix addition (SIMD Instructions)?

Question

I\'m trying to optimize 2d matrix addition in C using SIMD instructions (_mm256_add_pd, store, load, etc.). However, I\'m not seeing a large speedup at all. Using some timin