AVX2-based accelerated multiple precision floating-point programming with mixed precision iterative refinement - 42Papers