|
Size: 772
Comment:
|
Size: 1171
Comment:
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 1: | Line 1: |
| === Stuff I have done/been working on during Sage Dev 1 === | == Stuff I have done/been working on during Sage Dev 1 == |
| Line 3: | Line 3: |
| == LinBox 1.1.6 rc0 == | === LinBox 1.1.6 rc0 === |
| Line 7: | Line 7: |
| * fixing bugs,... | * fixing bugs in LinBox and Givaro,... * testing phase -> integration in sage-3.0.4 |
| Line 9: | Line 10: |
| == Dense Linear Algebra over GF(2) == * Design of the Matrix Multiply based Echelon decomposition |
=== Dense Linear Algebra over GF(2) === * Design of the Matrix Multiply based Echelon decomposition |
| Line 15: | Line 17: |
| == Benchmark(et)ing == | * LQUP (close to done * Still to be done * extensive testing of some routines * trtri (triangular inversion) * trtrm (triangular triangular matmul) === Algorithms for matrices overs Z and Zp[X] === * Theory: discussions and new insights with A Storjohann * practice: new libraries === Benchmark(et)ing === |
| Line 17: | Line 29: |
| * lobbying for introduction of Tbops for benchmaring M4RI | * lobbying for introduction of Tbops for benchmark**ing M4RI -> 1.2Tbops |
| Line 19: | Line 31: |
| == Playing with words == | === Playing with words === |
Stuff I have done/been working on during Sage Dev 1
LinBox 1.1.6 rc0
tried several ways of getting rid of gmp++ (required for debianization). ==> rely on Givaro's
integrated Sage's linbox_wrap interface in the LinBox library
fixing bugs in LinBox and Givaro,...
testing phase -> integration in sage-3.0.4
Dense Linear Algebra over GF(2)
* Design of the Matrix Multiply based Echelon decomposition
- learned a lot on the design and tricks of M4RI
- already done
- TSRM,
- generalization of matrix windows and algorithms working on them, to any dimensions
- LQUP (close to done
- Still to be done
- extensive testing of some routines
- trtri (triangular inversion)
- trtrm (triangular triangular matmul)
Algorithms for matrices overs Z and Zp[X]
- Theory: discussions and new insights with A Storjohann
- practice: new libraries
Benchmark(et)ing
- parallel ATLAS and fflas-ffpack: 70 Gffops
lobbying for introduction of Tbops for benchmark**ing M4RI -> 1.2Tbops
Playing with words
- some nicer prononciations for weird acronyms: M4RI and eMPIRe