HPF and numerical libraries. (English) Zbl 0931.65030
Zinterhof, Peter (ed.) et al., Parallel computation. 4th international ACPC conference including special tracks on Parallel numerics (ParNum ’99) and parallel computing in image processing, video processing, and multimedia. Salzburg, Austria, February 16–18, 1999. Proceedings. Berlin: Springer. Lect. Notes Comput. Sci. 1557, 140-152 (1999).
Summary: Portable and efficient ways for calling numerical high performance software libraries from high performance Fortran (HPF) programs are investigated. The methods suggested utilize HPF’s EXTRINSIC mechanism and are independent of implementation details of HPF compilers. Two prototypical examples are used to illustrate these techniques. Highly optimized BLAS routines are utilized for local computations: (i) in parallel multiplication of matrices, and (ii) in parallel Cholesky factorization. Both implementations turn out to be very efficient and show significant improvements over standard HPF implementations.
For the entire collection see [Zbl 0909.00047].
For the entire collection see [Zbl 0909.00047].
MSC:
| 65F05 | Direct numerical methods for linear systems and matrix inversion |
| 65Y05 | Parallel numerical computation |
| 65Y15 | Packaged methods for numerical algorithms |
