Go 1.27's SIMD ties with NumPy until the data fits in cache

Go 1.27 introduced an experimental simd package in the standard library. Performance tests against NumPy using a speaker-search index revealed varying results based on data size. With a 253 MB corpus, Go SIMD and NumPy (OpenBLAS) performed similarly, resulting in a tie. However, when using a 31 MB corpus that fits within the L3 cache, NumPy (OpenBLAS) proved significantly faster, outperforming Go SIMD by 2.4 times in median execution speed during the measured task.