2669
Comment:
|
5073
added wrap-up notes
|
Deletions are marked like this. | Additions are marked like this. |
Line 3: | Line 3: |
List any and all project ideas here. Use the format * (name) Project idea. If the idea is more than one line, make a [:/example: link like this (note the leading slash) to more info]. |
== Wrap-Up == * Benchmarking (was, malb, wjp) * developed a file format for benchmarks * wrote code which parses, runs, generates plots for these * going to be in the 2.9 release * might form the basis for new benchmarking document in the spirit of the constructions document |
Line 6: | Line 10: |
== Here they are == | * Cython Syntax Highlighting (robertwb) * visualization to show how good your Cython code is * HTML output |
Line 8: | Line 14: |
(If this list gets much longer -- which it will - we might want to have subheadings, e.g. number fields, linear algebra, elliptic curves... -- john.cremona) | * Wrapping g0n (jec) * g0n is optional package now * first wrapper written * memleaks fixed in g0n * Rewrapping Simon's Ell GP Scripts (robertwb, jec) * much simpler now because Simon updated code * Flint (Bill Hart) * many bugs fixed * Coercion (robertwb, roed, ...) * some beta available, other's can start working with it * Programming Guide * see [:days6/sprint/prog_guide: new plan] * FFT Improvements for NTL (dmharvey) * much faster on Opterons * not so much faster on Core2Duo (~ 20%) * available on [http://www.math.harvard.edu/~dmharvey/code/ntl-fft/index.html David's website] * automatic tuning required, not ready for inclusion * Sympy & SAGE (Ondrej) * mixing SAGE/SymPy expressions * Debian package of Cython (Ondrej) * preliminary package available * Matrix Permanents (Jaap) * hope to achieve 30 - 50% speed-up for permanents * Gröbner Basis Update (malb) * generic implementation for Gröbner bases over any field (slow!!) * implementation for Gröbner bases over ZZ (slow!!) * Doctest Benchmark (wjp) * timings of all doctests * basic comparison possible eg. 2.8.12 vs. 2.8.13 * PolyBoRi (Michael, Burcin) * Python Interface of PolyBoRi reimplemented in Cython * new wrapper can run most testcases * PolyBoRi release next weekend == Projects == * elliptic curves (John, rpw, was) * getting g0n into Sage * sparse linear algebra over $F_q$ * $\#E(F_q)$ * multivariate polynomial quotient rings (Burcin, malb, Ludovic, Michael Brickenstein) * generic polynomial division/reduction * quotient rings via libSINGULAR * PolyBoRi integration * FGB integration * coercion model/calculus/programming manual (robertwb, roed, John Hanke, Ondrey, David Loeffler, Jaap) * packages/updates/LinBox (cpernet, mabshoff) * LinBox bugfixes and updates * Givaro * IML * benchmarketing (robertwb, Willem) * doctest benchmarking interface == Originally Proposed Projects == |
Line 11: | Line 85: |
* (robertwb, roed, willem, jaap, jon hanke, david loeffler, ondrej) [:/prog_guide: Programming Guide] * (robertwb, roed, willem, jaap, jon hanke, david loeffler, ondrej) [:/coercion: Coercion] * (robertwb, roed, willem, jaap, jon hanke, david loeffler, ondrej) [:/calculus: Calculus] |
Sage Days 6 Coding Sprints
Wrap-Up
- Benchmarking (was, malb, wjp)
- developed a file format for benchmarks
- wrote code which parses, runs, generates plots for these
- going to be in the 2.9 release
- might form the basis for new benchmarking document in the spirit of the constructions document
- Cython Syntax Highlighting (robertwb)
- visualization to show how good your Cython code is
- HTML output
- Wrapping g0n (jec)
- g0n is optional package now
- first wrapper written
- memleaks fixed in g0n
- Rewrapping Simon's Ell GP Scripts (robertwb, jec)
- much simpler now because Simon updated code
- Flint (Bill Hart)
- many bugs fixed
- Coercion (robertwb, roed, ...)
- some beta available, other's can start working with it
- Programming Guide
- see [:days6/sprint/prog_guide: new plan]
- FFT Improvements for NTL (dmharvey)
- much faster on Opterons
not so much faster on Core2Duo (~ 20%)
available on [http://www.math.harvard.edu/~dmharvey/code/ntl-fft/index.html David's website]
- automatic tuning required, not ready for inclusion
Sympy & SAGE (Ondrej)
- mixing SAGE/SymPy expressions
- Debian package of Cython (Ondrej)
- preliminary package available
- Matrix Permanents (Jaap)
- hope to achieve 30 - 50% speed-up for permanents
- Gröbner Basis Update (malb)
- generic implementation for Gröbner bases over any field (slow!!)
- implementation for Gröbner bases over ZZ (slow!!)
- Doctest Benchmark (wjp)
- timings of all doctests
- basic comparison possible eg. 2.8.12 vs. 2.8.13
PolyBoRi (Michael, Burcin)
Projects
- elliptic curves (John, rpw, was)
- getting g0n into Sage
sparse linear algebra over F_q
\#E(F_q)
- multivariate polynomial quotient rings (Burcin, malb, Ludovic, Michael Brickenstein)
- generic polynomial division/reduction
- quotient rings via libSINGULAR
PolyBoRi integration
- FGB integration
- coercion model/calculus/programming manual (robertwb, roed, John Hanke, Ondrey, David Loeffler, Jaap)
- packages/updates/LinBox (cpernet, mabshoff)
LinBox bugfixes and updates
- Givaro
- IML
- benchmarketing (robertwb, Willem)
- doctest benchmarking interface
Originally Proposed Projects
- (wstein and ondrej certik) [:/debian: determine exactly which sage packages are in debian]
- (robertwb, roed, willem, jaap, jon hanke, david loeffler, ondrej) [:/prog_guide: Programming Guide]
- (robertwb, roed, willem, jaap, jon hanke, david loeffler, ondrej) [:/coercion: Coercion]
- (robertwb, roed, willem, jaap, jon hanke, david loeffler, ondrej) [:/calculus: Calculus]
- (wstein) upgrade the freetype package in sage.
- (wstein) upgrade the cvxopt package to 0.9
- (wstein) Number fields -- Reduction modulo a prime ideal, or any ideal.
- (wstein) Number fields -- optimize K([1,2,3,4,5]), etc., for K number field or order.
- (wstein) Timing of doctests (partially done)
- (wstein) Double question mark for spyx files doesn't work.; docstrings work.
- (wstein) Markup Cython code based on what C code it generates.
- (wstein) R Python expect interface
- (wstein) [:/edwards: Redo all the "edwards coord" proofs in Sage instead of Magma.]
- (wstein) ATLAS into Sage on Linux.
- (wstein) Linbox -- get rid of all the crappy hacks for minpoly/charpoly use in Sage.
- (wstein) Linbox -- fix wrapping of smith_form problems in Sage.
- (wstein) Create pickle database
- (wstein) Document number fields.
(wstein) Compute kernels of matrices over \mathbf{Z}/p^n\mathbf{Z}.
- (Ondrej Certik) simplify/improve calculus in SAGE [:/calculus: see this link for more info]
(mabshoff) fix BlasMatrix::Hom leak in LinBox
- (mabshoff) start of 2.8.13 release cycle
- (david harvey) try to improve cache efficiency of NTL's FFT routine
- (Bill Allombert) improve PARI ECM implementation
- (john.cremona) adapt interface to new Simon gp scripts
- (john.cremona) elliptic curve isomorphisms, urst transforms of curves and points
(john.cremona) improved \#E(F_q^d) for elliptic curves defined over F_q
(malb) speed up GF(p^n) for p \neq 2 and p^n \geq 2^{16} (using NTL::ZZ_pE and NTL::zz_pE)
- (malb) implement quotient rings over multivariate polynomial rings over finite fields using libSINGULAR
- (malb) implement multivariate polynomial rings over number fields using libSINGULAR
(malb) implement sparse matrices over GF(p^n) (using Givaro) modelled after sparse matrices modint.
- (weinmann) related to malb's point directly above: merge smatlib code into SAGE.
- (weinmann) extend MPolynomialSystemGenerator such that equation systems for (stream and block) ciphers can be generated in a generic way.