A BLAS-3 version of the QR factorization with column pivoting
- Univ. Jaime I, Castellon (Spain). Dept. de Informatica
- Duke Univ., Durham, NC (United States). Dept. of Computer Science
- Argonne National Lab., IL (United States). Mathematics and Computer Science Div.
The QR factorization with column pivoting (QRP), originally suggested by Golub is a popular approach to computing rank-revealing factorizations. Using Level 1 BLAS, it was implemented in LINPACK, and, using Level 2 BLAS, in LAPACK. While the Level 2 BLAS version delivers superior performance in general, it may result in worse performance for large matrix sizes due to cache effects. The authors introduce a modification of the QRP algorithm which allows the use of Level 3 BLAs kernels while maintaining the numerical behavior of the LINPACK and LAPACK implementations. Experimental comparisons of this approach with the LINPACK and LAPACK implementations on IBM RS/6000, SGI R8000, and DEC AXP platforms show considerable performance improvements.
- Research Organization:
- Argonne National Laboratory (ANL), Argonne, IL
- Sponsoring Organization:
- Comision Interministerial de Ciencia y Tecnologia (Spain); National Science Foundation, Washington, DC (United States); USDOE, Washington, DC (United States)
- DOE Contract Number:
- W-31109-ENG-38
- OSTI ID:
- 320999
- Journal Information:
- SIAM Journal on Scientific Computing, Journal Name: SIAM Journal on Scientific Computing Journal Issue: 5 Vol. 19; ISSN 1064-8275; ISSN SJOCE3
- Country of Publication:
- United States
- Language:
- English
Similar Records
Automatic Blocking Of QR and LU Factorizations for Locality
A block QR factorization algorithm using restricted pivoting