ParentFull threadlokopodium·Generic code yields better results than SSE/AVX optimized ones. I wonder why that could be.View on HN