| 1 | [[PageOutline]] |
| 2 | = Basic Linear Algebra Subroutines (BLAS) = |
| 3 | [http://www.netlib.org/blas/faq.html] |
| 4 | |
| 5 | Core routines used by LAPACK (Linear Algebra Package) and elsewhere. |
| 6 | Generally optimized for particular machine architectures, cache hierarchy. |
| 7 | |
| 8 | The subroutines are categorized to three levels. |
| 9 | * Level 1: Scalar and vector operations |
| 10 | * Level 2: Matrix-vector operations |
| 11 | * Level 3: Matrix-matrix operations |
| 12 | |
| 13 | Subroutine names start with: |
| 14 | * S: single precision |
| 15 | * D: double precision |
| 16 | * C: single precision complex |
| 17 | * Z: double precision complex |
| 18 | |
| 19 | Examples: |
| 20 | * DDOT: dot product of two vectors |
| 21 | * DGEMV: matrix-vector multiply, general matrices |
| 22 | * DGEMM: matrix-matrix multiply, general matrices |
| 23 | * DSYMM: matrix-matrix multiply, symmetric matrices |
| 24 | |
| 25 | List of Subroutine : [http://www.netlib.org/blas/blasqr.pdf] |
| 26 | |
| 27 | == Example == |
| 28 | * [[Blas Level 1]] |
| 29 | * [[Blas Level 2]] |
| 30 | * [[Blas Level 3]] |