Opened 6 years ago
Closed 6 years ago
#21778 closed defect (fixed)
correct refinement in cvxopt sdp backend
Reported by:  Dima Pasechnik  Owned by:  

Priority:  major  Milestone:  sage7.5 
Component:  numerical  Keywords:  
Cc:  Matthias Köppe, François Bissey  Merged in:  
Authors:  Dima Pasechnik  Reviewers:  Matthias Koeppe, François Bissey 
Report Upstream:  N/A  Work issues:  
Branch:  15690f5 (Commits, GitHub, GitLab)  Commit:  15690f53cadb93290e5c8db5ce9d063d20389032 
Dependencies:  Stopgaps: 
Description
on #20506 a weird numerical instability was noticed. It turned out to be a wrongly set solver parameter.
Change History (14)
comment:1 Changed 6 years ago by
Authors:  → Dima Pasechnik 

Branch:  → u/dimpase/t21778 
Cc:  Matthias Köppe added 
Commit:  → 4b6e69bf85ae8acc7892c1fde4620a0b5ee9e5e3 
Status:  new → needs_review 
comment:2 Changed 6 years ago by
On Mac OS X:
sage t warnlong 77.7 src/sage/numerical/sdp.pyx ********************************************************************** File "src/sage/numerical/sdp.pyx", line 97, in sage.numerical.sdp Failed example: p.dual_variable(1) # rel tol 1e05 Expected: [ 85555.0 85555.0] [85555.0 85555.0] Got: [ 85527.39489238134 85527.39489234997] [85527.39489234997 85527.39489238121] Tolerance exceeded in 4 of 4: 85555.0 vs 85527.39489238134, tolerance 3e04 > 1e05 85555.0 vs 85527.39489234997, tolerance 3e04 > 1e05 85555.0 vs 85527.39489234997, tolerance 3e04 > 1e05 85555.0 vs 85527.39489238121, tolerance 3e04 > 1e05 **********************************************************************
comment:3 Changed 6 years ago by
Commit:  4b6e69bf85ae8acc7892c1fde4620a0b5ee9e5e3 → 2dc7460857a4ff8f22cb7ea93e72509a3c198a0e 

Branch pushed to git repo; I updated commit sha1. New commits:
2dc7460  more tolerance

comment:5 Changed 6 years ago by
Cc:  François Bissey added 

Reviewers:  → Matthias Koeppe 
Works on Mac OS X now.
Cc'ing fbissey, who did some testing of this or related testcases on #20506.
comment:7 Changed 6 years ago by
Sorry, was that the one where I had to raise the tolerance when upgrading glpk
?
comment:8 Changed 6 years ago by
No, it is a different issue, with cvxopt interface that was broken and fixed by this patch.
comment:9 Changed 6 years ago by
OK now that I have the refresher I know what to do. to review this.
comment:10 Changed 6 years ago by
Reviewers:  Matthias Koeppe → Matthias Koeppe, François Bissey 

Status:  needs_review → needs_work 
OK the example of #20506 does work now. But I have a tolerance problem
sage t long /usr/lib64/python2.7/sitepackages/sage/numerical/sdp.pyx ********************************************************************** File "/usr/lib64/python2.7/sitepackages/sage/numerical/sdp.pyx", line 97, in sage.numerical.sdp Failed example: p.dual_variable(1) # rel tol 1e04 Expected: [ 85555.0 85555.0] [85555.0 85555.0] Got: [ 85527.39489238134 85527.39489234997] [85527.39489234997 85527.39489238121] Tolerance exceeded in 4 of 4: 85555.0 vs 85527.39489238134, tolerance 3e04 > 1e04 85555.0 vs 85527.39489234997, tolerance 3e04 > 1e04 85555.0 vs 85527.39489234997, tolerance 3e04 > 1e04 85555.0 vs 85527.39489238121, tolerance 3e04 > 1e04 **********************************************************************
comment:11 Changed 6 years ago by
Commit:  2dc7460857a4ff8f22cb7ea93e72509a3c198a0e → 15690f53cadb93290e5c8db5ce9d063d20389032 

comment:14 Changed 6 years ago by
Branch:  u/dimpase/t21778 → 15690f53cadb93290e5c8db5ce9d063d20389032 

Resolution:  → fixed 
Status:  positive_review → closed 
Note: See
TracTickets for help on using
tickets.
New commits:
Revert "removing numerically unstable test  will be on another ticket"
changed refinement to 1, adjusted tolerances