Differences between revisions 5 and 28 (spanning 23 versions)

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 new plan
FFT Improvements for NTL (dmharvey)
- much faster on Opterons
- not so much faster on Core2Duo (~ 20%)
- available on 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 Fq
- #E(Fq)
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
benchmarking (robertwb, Willem)
- doctest benchmarking interface

Originally Proposed Projects

(wstein and ondrej certik) determine exactly which sage packages are in debian
(robertwb, roed, willem, jaap, jon hanke, david loeffler, ondrej) Programming Guide
(robertwb, roed, willem, jaap, jon hanke, david loeffler, ondrej) Coercion
(robertwb, roed, willem, jaap, jon hanke, david loeffler, ondrej) 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) 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 Z/pnZ.
(Ondrej Certik) simplify/improve calculus in SAGE 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(Fqd) for elliptic curves defined over Fq
(malb) speed up GF(pn) for p/=2 and pn≥216 (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(pn) (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.

-  ⇤ ← Revision 5 as of 2007-11-11 15:23:14 → 
  Size: 882
  Editor: was2
  Comment:
+   ← Revision 28 as of 2010-10-29 04:29:46 → ⇥
  Size: 5106
  Editor: DanDrake
  Comment: remove link added by spammer
-Deletions are marked like this.
+Additions are marked like this.
 Line 3:
-List any and all project ideas here.
+== 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 5:
+Line 10:
- * Number fields -- Reduction modulo a prime ideal, or any ideal.
+ * Cython Syntax Highlighting (robertwb)
   * visualization to show how good your Cython code is
   * HTML output
-Line 7:
+Line 14:
- * Number fields --optimize K([1,2,3]).
+ * Wrapping g0n (jec)
   * g0n is optional package now
   * first wrapper written
   * memleaks fixed in g0n
-Line 9:
+Line 19:
- * Timing of doctests (partially done)
+ * Rewrapping Simon's Ell GP Scripts (robertwb, jec)
   * much simpler now because Simon updated code
-Line 11:
+Line 22:
- * Double question mark for spyx files doesn't work.; docstrings work.
+ * Flint (Bill Hart)
   * many bugs fixed
-Line 13:
+Line 25:
- * Markup Cython code based on what C code it generates.
+ * Coercion (robertwb, roed, ...)
   * some beta available, other's can start working with it
-Line 15:
+Line 28:
- * R Python expect interface
+ * Programming Guide
   * see [[days6/sprint/prog_guide| new plan]]
-Line 17:
+Line 31:
- * Redo all the "edwards coord" proofs in Sage instead of Magma.  See email to sagedays6, along with http://hyperelliptic.org/EFD/edwards.html
+ * 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
-Line 19:
+Line 37:
- * Fix computing trace over  $F_{q}$  algebraically.
+ * Sympy & SAGE (Ondrej)
   * mixing SAGE/SymPy expressions
  
 * Debian package of Cython (Ondrej)
   * preliminary package available
-Line 21:
+Line 43:
- * ATLAS into Sage on Linux.
+ * Matrix Permanents (Jaap)
   * hope to achieve 30 - 50% speed-up for permanents
-Line 23:
+Line 46:
- * Linbox -- get rid of all the crappy hacks for minpoly/charpoly
+ * Gröbner Basis Update (malb)
   * generic implementation for Gröbner bases over any field (slow!!)
   * implementation for Gröbner bases over ZZ (slow!!)
-Line 25:
+Line 50:
- * Linbox -- fix wrapping of smith_form problems.
+ * Doctest Benchmark (wjp)
   * timings of all doctests
   * basic comparison possible eg. 2.8.12 vs. 2.8.13
-Line 27:
+Line 54:
- * Create pickle database
+ * PolyBoRi (Michael, Burcin)
   * Python Interface of PolyBoRi reimplemented in Cython
   * new wrapper can run most testcases
   * PolyBoRi release next weekend
-Line 29:
+Line 59:
- * Document number fields.
+== Projects ==
 * elliptic curves (John, rpw, was)
  * getting g0n into Sage
  * sparse linear algebra over  $F_{q}$ 
  *  $# E (F_{q})$
-Line 31:
+Line 65:
- * Compute kernels of matrices over  $Z / p^{n} Z$ .
+ * 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)

 * [[days6/sprint/packages|packages/updates/LinBox]] (cpernet, mabshoff)
  * LinBox bugfixes and updates
  * Givaro
  * IML

 * benchmarking (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  $Z / p^{n} Z$ .

 * (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  $G F (p^{n})$  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  $G F (p^{n})$  (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.

Diff for "days6/sprint"

Sage Days 6 Coding Sprints

Wrap-Up

Projects

Originally Proposed Projects