Acceleration of LU decomposition supporting double-double, triple-double, and quadruple-double precision floating-point arithmetic with AVX2 | IEEE Conference Publication | IEEE Xplore