Skip to content

Latest commit

 

History

History
190 lines (130 loc) · 7.55 KB

Changelog.md

File metadata and controls

190 lines (130 loc) · 7.55 KB
Raw

SeDuMi 1.3.7 (2023-03-18)

https://github.com/sqlp/sedumi/compare/v1.3.6...v1.3.7

SeDuMi 1.3.6 (2023-03-17)

https://github.com/sqlp/sedumi/compare/v1.3.5...v1.3.6

Many thanks to @araujoms for his major contributions to this release.

SeDuMi 1.3.5 (2021-06-15)

https://github.com/sqlp/sedumi/compare/v1.3.4...v1.3.5

SeDuMi 1.3.4 (2020-01-09)

https://github.com/sqlp/sedumi/compare/v1.3.2...v1.3.4

Note: version 1.3.3 was not officially released.

  • Many many bug fixes in seven years.

SeDuMi 1.3.2 (2013-07-24, Michael Grant)

https://github.com/sqlp/sedumi/compare/v1.3.1...v1.3.2

  • Merged changes from CVX and SeDuMi/Octave into the distribution.
  • Further changes to suppress Matlab Code Analyzer warnings.
  • Removed matrix.h dependency from givens.h for Octave compatibility; fixed assertion problem with givensrot.c.

SeDuMi 1.3.1 (2013-04-04)

https://github.com/sqlp/sedumi/compare/v1.3.0...v1.3.1

  • Corrected bugs in sedumi.m and getada.m.
  • Updated .m files as per mlint suggestions.
  • Commented length(b) vs length(c) check in pretransfo.
  • Recompiled windows mex files, updated install_sedumi.m for MATLAB v8

New features in SeDuMi 1.1 since 1.05R5

  • 2010-02-15 Fix: global variables are used for some large data structures (ADA, At).

  • 2010-02-15 Fix: getada1 and getada2 have been backported to Matlab, which actually increases speed, due to parallel BLAS.

  • 2009-11-25 Fix: computation in getsymbada.m has been reorganized for speed.

  • 2009-11-15 Fix: the psdeig, psdinvscale, psdjmulv, psdfactor, psdscale and triumtriu mex files were replaced with a pure Matlab equivalent.

  • 2009-09-05 Fix: SeDuMi now uses wall clock time instead of cputime, as the CPU time is misleading in parallel processing environments. It is reported in info.walltime. info.cpusec contains the total cpu time, which may be more than the wall time.

  • 2009-08-05 Fix: Based on feedback from Didier Henrion and others, the rank and infeasibility detection code has been modified.

  • 2009-05-10 Fix: Michael C. Grant modified some files to compile under 64-bit versions.

  • 2009-05-10 Fix: A new installer script is provided by Michael C. Grant.

  • 2009-05-10 Fix: Some BLAS calls are fixed in blkaux.c.

  • 2008-04-09 Fix: BLAS is used to improve performance.

  • 2008-04-09 Fix: SeDuMi now works on 64-bit operating systems.

  • 2006-10-14 Fix: A bugfix was contributed by Michael C. Grant regarding the handling of 2-dimensional Lorentz cones.

  • 2006-10-10 Fix: A new, completely rewritten SDPA format reader is provided in the conversion folder. It is fast and it follows the SDPA standard.

  • 2006-10-10 Fix: Michael Grant corrected a bug in the postprocessing phase.

  • 2006-10-10 Fix: The semidefinite preprocessing assumed that there are no free variables.

  • 2006-10-10 Fix: Matlab R2006b does not support fprintf with fid=0, (i.e., no output). A workaround is provided. Thanks to Johan Löfberg.

  • 2006-10-10 Fix: The solutions were incorrect if the error measures were turned on.

  • 2006-07-12 Fix: A bug was discovered by Paul Tseng when using free variables together with rotated Lorentz cones.

  • 2005-06-22 Fix: A bug about free variable handling has been reported by Johan Löfberg, the fixed version is now posted.

  • 2005-06-21 Fix: A bug about complex variables has been reported by Michael C. Grant, the correct version is now available.

  • 2005-06-10 Fix: && and || were replaced by & and | to make the package compatible with Matlab 6.0 (R12).

  • 2005-05-10 Feature: The binaries are now built from Matlab, see Install.txt for details.

  • 2005-03-02 Fix: The 'See also' references at the end of the help portions are now clickable.

  • 2005-03-02 Fix: If ADA is actually dense then the symbolic Cholesky factorization and the minimum degree ordering are not performed. Also in this case SYMBADA is created directly as a fully dense matrix of ones stored as sparse.

  • 2005-03-02 Feature: Included a simple heuristic to turn step-differentiation on/off.

  • 2005-03-02 Feature: New options to control preprocessing: pars.free, pars.sdp.

  • 2005-03-02 Feature: Preprocessing routines are now included. Diagonal SDP blocks are converted into nonnegative variables. Free variables are handled in a quadratic cone, split free variables are detected in the linear part. The sparsity structure of A is corrected at the end of preprocessing.

  • 2005-03-02 Fix: New default values. pars.eps = 1e-8, pars.stepdif = 2, pars.cg.qprec = 1.

  • 2005-02-02 Feature: If pars.errors=1 then SeDuMi outputs the six errors measures defined in the 7th DIMACS challenge paper that is available at http://plato.asu.edu/dimacs/node3.html. These are in info.err1, ..., info.err6.

  • 2005-01-20 Fix: Minor speed improvement in the sdmaux* files (loop unrolling).

  • 2005-01-10 Fix: Minor speed improvement in popK.m.

  • 2005-01-10 Fix: blkchol is now invoked directly from SeDuMi.m without sparchol.

  • 2005-01-10 Fix: A small change in Amul.m resulted in drastic speed improvement for middle-sized problems. Additionally, sum(c.*x) is much faster than c'*x if c and x are sparse. The moral is that taking the transpose of a sparse matrix is slow.

  • 2004-01-03 Fix: Fixed a small bug in wregion.m concerning division by zero. (Johan Löfberg)

  • 2004-11-29 Fix: Many if-then statements checking for data validity in the C code were replaced by mxAssert statements. The only exceptions are memory failure checks.

  • 2004-11-29 Feature: Timing is changed. So far SeDuMi reported the CPU time spent in the main IPM loop. Now info.timing is a vector containing the time spent in the preprocessing, IPM and postprocessing, respectively. The old info.cpusec contains the total time of the algorithm. This does not make much difference now, since 99% of the time is spent in the IPM loop, but if more stuff is added to the pre- and postprocessing, this can be an issue.

  • 2004-11-11 Fix: Cleaned up the code with M-Lint. All unused variables and never used assignments are removed. All | and & are replaced with || and && whenever it was possible. This speeds up conditional statements in the preprocessing phase. Unused .m and .c files are removed from the distribution.

  • 2004-11-03 Feature: pars.stopat can now be a vector. The algorithm enters debug mode at each iteration specified in pars.stopat. The order does not matter, invalid values are ignored.