Skip to content

Use configured feasibility tolerance in Papilo #1688

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
rapids-bot merged 4 commits into from
Aug 11, 2026
Merged

Use configured feasibility tolerance in Papilo #1688

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
c6ef256
Use configured feasibility tolerance in PaPILO
hlinsen Aug 6, 2026
5002b48
Merge branch 'main' of github.com:NVIDIA/cuopt into fix-papilo-feastol
hlinsen Aug 10, 2026
1e5f376
Fix merge conflicts
hlinsen Aug 10, 2026
70beaf7
Merge branch 'main' of github.com:NVIDIA/cuopt into fix-papilo-feastol
hlinsen Aug 11, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions cpp/src/math_optimization/solver_settings.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -94,8 +94,8 @@ solver_settings_t<i_t, f_t>::solver_settings_t() : pdlp_settings(), mip_settings
{CUOPT_RELATIVE_PRIMAL_TOLERANCE, &pdlp_settings.tolerances.relative_primal_tolerance, f_t(0.0), f_t(1e-1), f_t(1e-4)},
{CUOPT_ABSOLUTE_GAP_TOLERANCE, &pdlp_settings.tolerances.absolute_gap_tolerance, f_t(0.0), f_t(1e-1), f_t(1e-4)},
{CUOPT_RELATIVE_GAP_TOLERANCE, &pdlp_settings.tolerances.relative_gap_tolerance, f_t(0.0), f_t(1e-1), f_t(1e-4)},
{CUOPT_MIP_ABSOLUTE_TOLERANCE, &mip_settings.tolerances.absolute_tolerance, f_t(0.0), f_t(1e-1), f_t(1e-4)},
{CUOPT_MIP_RELATIVE_TOLERANCE, &mip_settings.tolerances.relative_tolerance, f_t(0.0), f_t(1e-1), f_t(1e-4)},
{CUOPT_MIP_ABSOLUTE_TOLERANCE, &mip_settings.tolerances.absolute_tolerance, f_t(0.0), f_t(1e-1), f_t(1e-6)},
{CUOPT_MIP_RELATIVE_TOLERANCE, &mip_settings.tolerances.relative_tolerance, f_t(0.0), f_t(1e-1), f_t(1e-12)},
{CUOPT_MIP_INTEGRALITY_TOLERANCE, &mip_settings.tolerances.integrality_tolerance, f_t(0.0), f_t(1e-1), f_t(1e-5)},
{CUOPT_MIP_ABSOLUTE_GAP, &mip_settings.tolerances.absolute_mip_gap, f_t(0.0), std::numeric_limits<f_t>::infinity(), std::max(f_t(1e-10), std::numeric_limits<f_t>::epsilon())},
{CUOPT_MIP_RELATIVE_GAP, &mip_settings.tolerances.relative_mip_gap, f_t(0.0), f_t(1e-1), f_t(1e-4)},
Expand Down
2 changes: 1 addition & 1 deletion cpp/src/mip_heuristics/presolve/third_party_presolve.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -722,7 +722,7 @@ void set_presolve_options(papilo::Presolve<f_t>& presolver,
{
presolver.getPresolveOptions().tlim = time_limit;
presolver.getPresolveOptions().threads = num_cpu_threads; // user setting or 0 (automatic)
presolver.getPresolveOptions().feastol = 1e-5;
presolver.getPresolveOptions().feastol = absolute_tolerance;
if (dual_postsolve) {
presolver.getPresolveOptions().componentsmaxint = -1;
presolver.getPresolveOptions().detectlindep = 0;
Expand Down