Several bug fixes

[nguidotti]

• Fixed early termination in CMS750_4 due to an incorrect root objective from the PDLP. It provides an objective in user space and B&B expects the objective in the solver space.
• Fixed setting the number of threads to a fixed value when running cuopt_cli. Now the number of threads in the probing cache can be controlled by the bounds_update::settings_t.
• Inverted loop order in probing cache to avoid creating and deleting the OpenMP thread pool each iteration.
• PDLP/Barrier no longer set the root objective and solution when the B&B is already running.
• Improved the logs when solving the root relaxation using the concurrent mode.

Checklist

• I am familiar with the Contributing Guidelines.
• Testing
  • New or existing tests cover these changes
  • Added tests
  • Created an issue to follow-up
  • NA
• Documentation
  • The documentation is up to date with these changes
  • Added new documentation
  • NA

Summary by CodeRabbit

• New Features

  • Per-run threading control and explicit bidirectional conversion between user-space and solver-space objectives.

• Optimizations

  • Presolve and preprocessing use configurable thread counts to align concurrency with instance settings.

• Bug Fixes

  • Prevents unintended root-solution overrides and stabilizes root handling across concurrent solves.

• Behavior Changes

  • Root-solution callbacks now receive solver timing plus both solver-space and user-space objective values.

• Quality of Life

  • Reduced verbose solver logging during mixed-mode MIP runs.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[coderabbitai]

📝 Walkthrough

Walkthrough

Treats PDLP objective as a user-space value and adds a user→solver objective conversion API; passes solver- and user-space objectives plus solve time to B&B callbacks; makes presolve threading configurable with OpenMP; records which solver solved the root; and suppresses LP solver logs during MIP solves via a conditional logging macro.

Changes

Cohort / File(s)	Summary
Objective conversion & callback API cpp/src/mip/problem/problem.cuh, cpp/src/mip/problem/problem.cu	Adds f_t get_solver_obj_from_user_obj(f_t user_obj) const to problem_t and updates set_root_relaxation_solution_callback type to accept an additional f_t so callbacks receive solver-space objective and/or solve_time alongside the user objective.
Diversity manager objective handling cpp/src/mip/diversity/diversity_manager.cu	Treats PDLP objective as a user-space value, derives solver_obj via the new API, retrieves solve_time, and updates the B&B callback invocation to pass the solver-space objective, user-space objective, and solve_time with adjusted argument ordering and clarifying comments.
Branch-and-bound root tracking cpp/src/dual_simplex/branch_and_bound.hpp, cpp/src/dual_simplex/branch_and_bound.cpp, cpp/src/mip/solver.cu	Adds enum class root_solver_type_t { NONE, CROSSOVER, DUAL_SIMPLEX } and std::atomic<root_solver_type_>; changes set_root_relaxation_solution to accept solve_time, record root solver type, and populate root data only once; updates B&B callback binding to include the extra placeholder.
Presolve threading / probing cache cpp/src/mip/presolve/bounds_presolve.cuh, cpp/src/mip/presolve/probing_cache.cu	Adds i_t num_threads = -1 setting and #include <omp.h>; computes/clamps num_threads, sizes per-thread pools by num_threads, and uses #pragma omp parallel num_threads(num_threads) to bind concurrency to the configured value.
LP solver logging cpp/src/linear_programming/solve.cu	Introduces CUOPT_LOG_CONDITIONAL_INFO(condition, ...) macro and replaces several CUOPT_LOG_INFO calls with it to suppress verbose solver logs inside MIP solves; updates copyright year.
Build manifest CMakeLists.txt	Small CMake adjustments (few lines changed).

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

🚥 Pre-merge checks | ✅ 1 | ❌ 2

❌ Failed checks (1 warning, 1 inconclusive)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 10.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.
Title check	❓ Inconclusive	The title 'Several bug fixes' is vague and generic. While the PR does address multiple bug fixes, the title uses non-descriptive language that fails to convey the specific nature or scope of the changes to someone scanning the commit history.	Consider using a more specific title that highlights the primary bug fix, such as 'Fix PDLP objective space mapping in B&B root relaxation' or naming the most critical issue addressed.

✅ Passed checks (1 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

• 📝 Generate docstrings

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 0

Caution

Some comments are outside the diff and can’t be posted inline due to platform limitations.

⚠️ Outside diff range comments (1)

cpp/src/mip/presolve/probing_cache.cu (1)

859-910: Add defensive validation for num_threads before pool sizing and parallel launch

The code assumes settings.num_threads is always positive, but since settings_t is a public struct, external code can set num_threads to 0 (or modify it between initialization and use). If 0 is assigned:

• modification_vector_pool and substitution_vector_pool would have size 0
• #pragma omp parallel num_threads(0) produces undefined behavior (OpenMP spec)
• If threads spawn despite 0, thread_idx from omp_get_thread_num() will exceed pool bounds, causing out-of-bounds access

Add a defensive clamp before sizing the pools:

Proposed fix

-  const size_t num_threads = bound_presolve.settings.num_threads;
+  const i_t requested_threads = bound_presolve.settings.num_threads;
+  size_t num_threads =
+    requested_threads > 0 ? static_cast<size_t>(requested_threads)
+                          : static_cast<size_t>(omp_get_max_threads());
+  if (num_threads == 0) { num_threads = 1; }

Aligns with multithreading safety guidelines to prevent resource exhaustion and thread safety violations.

[akifcorduk]

Can't we undef the previous macro and still use CUOPT_LOG_INFO here?

[nguidotti]

I think it is better to add another macro, since there are some methods where we do not pass the setting as argument, hence they need to use the old version.

[nguidotti]

Ideally, we should add logger_t to the settings rather than CUOPT_LOG_INFO or rapids::logger_t.

[chris-maes]

Let's move this printing to inside solve_root_relaxation and only print when crossover is successful.

Otherwise we could tell the user we have found a solution with PDLP/Barrier. And then the user could wait a long time only to have us use the dual simplex solution because we either: a) failed in crossover, or b) crossover took a very long time.

[chris-maes]

Do we need this? Can we just determine who won in solve_root_relaxation?

[chris-maes]

This is where we should indicate that PDLP/Barrier won. Can you use the is_pdlp_solution to determine which of them won and print it here?

[nguidotti]

I did not find a way to get the winner solver from the optimization_problem_solution_t unless we change the API (#787). This object is user-facing, so it needs to be reflect on the Python side as well.

[chris-maes]

I would remove this print. I don't think we want to interrupt the dual simplex logs for the root solve, unless we know that crossover was successful.

[hlinsen]

Does this need to be an atomic? I think we can never have a race condition because we write to this var only when one f the solve is optimal which means the other solver returned with halt status.