- Apr 12, 2013
-
-
Sébastien Villemot authored
-
Sébastien Villemot authored
-
- Apr 11, 2013
-
-
Houtan Bastani authored
-
Houtan Bastani authored
-
Sébastien Villemot authored
Closes #353
-
Sébastien Villemot authored
Closes #307
-
Sébastien Villemot authored
Fix issue about reduced form mapping with posterior samples + notation improvements
-
Marco Ratto authored
2) Improve the notation of info saved of threshold reduced form mapping which was obscure.
-
- Apr 10, 2013
-
-
Houtan Bastani authored
-
Houtan Bastani authored
-
Houtan Bastani authored
-
Houtan Bastani authored
-
Sébastien Villemot authored
Clarification of terminology of filtered variables
-
Johannes Pfeifer authored
Closes #317
-
- Apr 09, 2013
-
-
Houtan Bastani authored
-
Houtan Bastani authored
-
Houtan Bastani authored
-
MichelJuillard authored
power of negative numbers
-
Houtan Bastani authored
reporting: @graph: use green shading which is easier to see because 'FaceAlpha' doesn't work on octave (at least on os x)
-
Houtan Bastani authored
-
Houtan Bastani authored
-
- Apr 08, 2013
-
-
Sébastien Villemot authored
parallel bug fix
-
Marco Ratto authored
-
- Apr 05, 2013
-
-
Stéphane Adjemian authored
several bug fixes for cluster computations
-
Marco Ratto authored
1) properly manage hybrid matlab/octave computations; 2) check for function handles under octave, which cannot be saved on a mat file; 3) bug fix for non standard octave path; 4) synchronization fix under octave; 5) Retrieve missing output files when remote threads are finished; 6) check if input files have been updated, like in metropolis computations with load_mh_file=1; 7) check if output files are missing to trap synchronization failures: this may be inefficient and needs to be improved for unix/windows clusters; 8) fixed cleanup with multiple remote nodes.
-
Marco Ratto authored
-
Marco Ratto authored
-
Marco Ratto authored
2) removed return after failure of ping; some machines like hal can be connected but do not allow ping;
-
Marco Ratto authored
-
Marco Ratto authored
-
Marco Ratto authored
-
Sébastien Villemot authored
It is anyway impossible to have exactly the same behaviour of ilu and luinc with respect to zero filling and maximum pivot
-
Houtan Bastani authored
-
Houtan Bastani authored
-
- Apr 04, 2013
-
-
Houtan Bastani authored
-
Houtan Bastani authored
-
Houtan Bastani authored
-
Houtan Bastani authored
-
Houtan Bastani authored
-
- Apr 03, 2013
-
-
Houtan Bastani authored
-