In this paper, we propose a sparse approximate inverse for triangular matrices (SAIT) based on Jacobi iteration. The main operation of the algorithm is matrix-matrix multiplication. We apply the SAIT to iterative methods with ILU preconditioners. Then the two triangular solvers in the ILU preconditioning procedure are replaced by two matrix-vector multiplications, which can be fine-grained parallelized. We test the new algorithm by solving some linear systems and eigenvalue problems.
翻译:在本文中,我们提出了基于代谢的三角矩阵(SAIT)的微小近似反差。 算法的主要操作是矩阵矩阵矩阵乘法。 我们将SAIT应用于与ILU先决条件者一起的迭代方法。 然后, ILU 先决条件程序中的两个三角解答器被两个矩阵矢量乘法所取代, 这两个乘法可以细微地平行。 我们通过解决某些线性系统和电子价值问题来测试新的算法。