{"id":"arxiv/2610.30227","url":"https://arxiv.org/abs/2610.30227","tags":["math.NA"],"score":0,"title":"DBDSVDMR3: An MRRR-Based Bidiagonal SVD Solver Built from DSTEGR","domain":"math.NA","source":"arxiv","authors":["James Demmel","Inderjit Dhillon","Sai Surya Duvvuri","Julien Langou","Osni Marques","Ryan Schneider"],"reviews":0,"subject":"math.NA","summary":"The optimal algorithm for the symmetric tridiagonal eigenvalue problem is Multiple Relatively Robust Representations (MRRR or MR$^3$), which is capable of computing $k$ eigenvalue/eigenvector pairs in only $O(kn)$ floating-point operations. While MRRR is nearly 30 years old, it has never successfully been adapted to the bidiagonal SVD. This is not for lack of trying; two previous attempts produced implementations that were considered for LAPACK but eventually abandoned, a result of the fact that neither was reliable enough when run on test problems. The question therefore remained: is there…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-10-21","reactions":{}}