Abstract : This paper focuses on the performance analysis of a linear system solving based on Cholesky decomposition and QR factorization, implemented on 16bits fixed-point DSP-chip (TMS320C6474). The classical method of Cholesky decomposition has the advantage of low execution time. However, the modified Gram-Schmidt QR factorization performs better in term of robustness against the round-off error propagation. In this study, we have proposed a third method called Modified Gram-Schmidt Cholesky Decomposition. We have shown that it provides a compromise of the two performance criterias cited above. A joint theoretical and experimental analysis of global performance of the three methods has been presented and discussed.
https://hal-cnam.archives-ouvertes.fr/hal-02448976
Contributor : Christophe Alexandre
<>
Submitted on : Tuesday, February 11, 2020 - 12:57:10 PM
Last modification on : Thursday, February 13, 2020 - 11:05:02 AM
Rabah Maoudj, Luc Fety, Christophe Alexandre. Performance Analysis of Modified Gram-Schmidt Cholesky Implementation on 16 bits-DSP-chip. International Journal of Computing and Digital Systems, University of Bahrain, 2013, 2 (1), pp.21-27. ⟨10.12785/ijcds/020103⟩. ⟨hal-02448976⟩
