Multifrontral multithreaded rank-revealing sparse QR factorization

摘要

SuiteSparseQR is a sparse multifrontal QR factorization algorithm.Dense matrix methods within each frontal matrix enablethe method to obtain high performance on multicore architectures. Parallelismacross different frontal matrices is handled with Intelu27s Threading BuildingBlocks library.Rank-detection is performed within eachfrontal matrix using Heathu27s method, which does not require column pivoting.The resulting sparse QR factorization obtains a substantial fraction of thetheoretical peak performance of a multicore computer.