| | 1 | 2007-07-05 v1.2.2 |
| | 2 | * add a transposed version of the LQUP decomposition routine |
| | 3 | LUdivine |
| | 4 | * fix many bugs in LUdivine |
| | 5 | * new schedules for Winograd algorithm for matrix multiplication: |
| | 6 | 2 cases depending whether beta = 0 or not, taken form [Huss |
| | 7 | Ledermann & Al. 96] |
| | 8 | * add rowEchelon and ReducedRowEchelon routines + associated tests |
| | 9 | 2007-06-21 v1.2.1 |
| | 10 | * add the use of float BLAS, if the field caradinality is small enough |
| | 11 | * improve genericity: gemm can be use over any field domain (not |
| | 12 | requiring any conversion to a integral representation) |
| | 13 | * add a variant of Winograd's algorithm with less temporaries for |
| | 14 | the operation C = AxB |
| | 15 | * add ColumnEchelon and ReducedColumnEchelon routines, using an |
| | 16 | inplace algorithm, based on the LQUP decompositon of LUdivine |
| | 17 | * add routines ftrtri (replacing invL), ftrtrm. |
| | 18 | * fix bunch of memory leaks in the tests (not yet finished) |
