NVIDIA cuOpt Files

What's Changed

🚨 Breaking Changes

• build: don't install libcuopt into conda env by default by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1650
• Presolve work and iteration limits by @akifcorduk in https://github.com/NVIDIA/cuopt/pull/1656
• refactor: make to_optimization_problem a free function by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1802
• Migrate stream APIs from rmm::cuda_stream_view to cuda::stream_ref by @bdice in https://github.com/NVIDIA/cuopt/pull/1828
• build(cmake): split libcuopt into cuopt_client / cuopt_mathopt / cuopt_routing component libs by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1622
• build(cmake): give each component its own install component by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1926
• modify docker image to run proxy+grpc by default by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2007
• build(wheel): make libcuopt a thin metapackage by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2013

🐛 Bug Fixes

• GF2 presolve rank-deficiency fix by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1660
• Fix crash when symmetry generators are pruned by @nguidotti in https://github.com/NVIDIA/cuopt/pull/1689
• [BUG] Fixed uninitialized FJ constraint weights when the problem grows by @nguidotti in https://github.com/NVIDIA/cuopt/pull/1698
• Use configured feasibility tolerance in Papilo by @hlinsen in https://github.com/NVIDIA/cuopt/pull/1688
• Fix GF2 incorrectly treating some one-sided rows as equality by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1707
• Fix dual residual computation in final solution output by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/1718
• Remove split compile by default by @Kh4ster in https://github.com/NVIDIA/cuopt/pull/1722
• Report absent solution values instead of returning success by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1734
• fix: avoid uncapturable RMM copies inside routing CUDA graph captures by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1753
• fix: keep host sources alive for async route-view copies (#1756) by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1759
• Revert changes to PCG by @nguidotti in https://github.com/NVIDIA/cuopt/pull/1762
• Fix QCQP dual solution overrun and constraint count reporting by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/1761
• fix: build-images CI failures (deadsnakes PPA flakiness + nightly version mismatch) by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1780
• Fix mod2 integrality tolerance by @akifcorduk in https://github.com/NVIDIA/cuopt/pull/1770
• Fix cuopt_assert macro argument expansion by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1784
• Fix device_scalar constructors after RMM change by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1795
• ci: pass the build datetime through to image builds by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1792
• Correctly return reduced costs for PDLP (stable3) by @mlubin in https://github.com/NVIDIA/cuopt/pull/1797
• Add missing TransactionGuard in the SLDA presolver by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1815
• Fix cuopt_mvn retry loop being silently skipped under set -e by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1823
• Route Java console logging through System.out instead of native stdout by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1825
• Fix asynchronous clique table handoff by @hlinsen in https://github.com/NVIDIA/cuopt/pull/1834
• Disable clique/zero-half cuts in determinism mode by @nguidotti in https://github.com/NVIDIA/cuopt/pull/1841
• Forward routing gRPC settings and map node name strings to ints by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1837
• fix(build): pin RAPIDS_BRANCH and shared-workflows to release/26.10 by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1870
• change Python grpc client so wait loop happens in Python by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1857
• Concurrent halt fix on barrier by @akifcorduk in https://github.com/NVIDIA/cuopt/pull/1810
• fix(build): repair cuda::stream_ref migration breakage from [#1828] by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1894
• fix(docker): keep libgomp1 in the Ubuntu runtime image (cuDSS threading layer, [#1670]) by @vgong-nv in https://github.com/NVIDIA/cuopt/pull/1888
• ci(thirdparty-cvxpy): fix cwd import-shadowing that masked the installed wheel by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1891
• Routing: avoid division by zero when vehicle capacity is zero by @vitor1001 in https://github.com/NVIDIA/cuopt/pull/1853
• Guard against nullptr dereference when checking concurrent solver status by @vitor1001 in https://github.com/NVIDIA/cuopt/pull/1845
• Routing: Guard diagonal element lookup against negative offsets in top_k by @vitor1001 in https://github.com/NVIDIA/cuopt/pull/1855
• Routing: Fix vector lifetimes in breaks unit test by @vitor1001 in https://github.com/NVIDIA/cuopt/pull/1854
• docs: update NVIDIA Brev documentation links by @ZedingZhang in https://github.com/NVIDIA/cuopt/pull/1925
• Pin cuDSS's dlopen'd threading-layer plugin to prevent a teardown race by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1883
• Fix a few CI issues by @KyleFromNVIDIA in https://github.com/NVIDIA/cuopt/pull/1927
• presolve: refresh the cached linear column count after removing empty… by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/1923
• Fix issue in dual simplex where cut off used perturbed objective; use original objective and require dual feasibility by @chris-maes in https://github.com/NVIDIA/cuopt/pull/1901
• grpc server align ping handling with client behavior by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1937
• Fix run mip tolerance by @akifcorduk in https://github.com/NVIDIA/cuopt/pull/1956
• wire MIP initial solutions into gRPC by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1868
• [BUG] Correctly stop sub-MIP when main solve finishes by @nguidotti in https://github.com/NVIDIA/cuopt/pull/1911
• refactor(routing): stop cpu_routing_problem.hpp from leaking raft/rmm by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1969
• Fix Java JNI build on Ubuntu 22.04, add a Java smoke test to image CI by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1974
• Recompute dual variables and check dual feasibility after basis repair by @chris-maes in https://github.com/NVIDIA/cuopt/pull/1965
• fix(grpc): regenerate cuopt_mcp_schema.json, fail codegen drift earlier in CI by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1980
• proxy server make msgpack the default for wildcard ACCEPT by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1964
• Fix false optimality declaration in the determinsitic mode by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1954
• fix: escape parser error messages by @neuriv in https://github.com/NVIDIA/cuopt/pull/1947
• bugfix: handle cuda graph errors by @neuriv in https://github.com/NVIDIA/cuopt/pull/1945
• proxy: always include warnings and notes on result-file stubs by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1990
• proxy: DELETE of an unknown request returns 200 with zero counts by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1991
• proxy: say file vs stream in optimization-data 422s by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1992
• fix(ci/docker): manifest --amend duplicate entries + unused gdb-gdbserver by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1997
• proxy server match legacy behavior on missing logs by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1989
• track and return solve time for VRP solves in the gRPC server to support proxy server parity with legacy by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1985
• proxy: use a stable job 404 and stop UUID-gating job routes for parity with legacy by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1987
• Restore parameter logging by @nguidotti in https://github.com/NVIDIA/cuopt/pull/1996
• Bug fix: barrier determinism by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/1856
• Fix concurrent clique table publication by @derluke in https://github.com/NVIDIA/cuopt/pull/1775
• Fix solution queue memory overflow by @akifcorduk in https://github.com/NVIDIA/cuopt/pull/2002
• fix(routing): declare cusparse, which libcuopt_routing.so links by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2008
• fix(cmake): ship constants.h with the component that needs it by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2016
• [BUG] Fix delayed termination during root loop by @nguidotti in https://github.com/NVIDIA/cuopt/pull/2003
• [BUG] Fix long interval between progress report by @nguidotti in https://github.com/NVIDIA/cuopt/pull/2019
• Fix ranged constraint conversion and MPS round trips by @rg20 in https://github.com/NVIDIA/cuopt/pull/2027
• Fix erroneous early solve termination by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/2037
• Fix/mip time limit overruns by @akifcorduk in https://github.com/NVIDIA/cuopt/pull/2046
• grpc pdlp warmstart data was not transmitted and read by the server by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2041
• fix(wheel): declare libraft, whose headers the public API needs by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2014
• Reject unsupported mixed-integer quadratic problems before solving by @rg20 in https://github.com/NVIDIA/cuopt/pull/2032
• Bug fix: Insert barrier slack columns in front of the cone block for range rows by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/2029
• Fix MPS writing for empty CSR constraint matrices by @mlubin in https://github.com/NVIDIA/cuopt/pull/2039
• fix(ci): pin remaining shared-workflows refs to release/26.10 by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2040
• Fix knapsack rational coefficient scaling by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1754
• fix(test): update grpc_integration_test to submit_mip's new solver_settings_t signature by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2055
• fix(wheel): symlink component VERSION files to the repo-root VERSION by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2061
• fix(docker): build the Java JNI library against the split libcuopt wheels by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2062
• grpc server add a check for successful cuda init in worker by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2063
• Fix continuous variable elimination tolerance in MIP presolve by @chris-maes in https://github.com/NVIDIA/cuopt/pull/2066
• Add extra termination checks in LP and submip by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/2054
• fix(docker): install the unsuffixed nvidia-cuda-cccl for every CUDA major by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2075
• Dual simplex objective logging fixes by @chris-maes in https://github.com/NVIDIA/cuopt/pull/2072
• grpc and proxy server, return the bound value for incumbents by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2079
• skip job cancellation tests temporarily by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2081
• Fix scale fallback direction in dual simplex by @chris-maes in https://github.com/NVIDIA/cuopt/pull/2080
• Fix FJ score alignment for CUDA atomic loads by @mlubin in https://github.com/NVIDIA/cuopt/pull/2088
• fix(linear_programming): implement neg, pos, and abs for LinearExpression (fixes [#2056]) by @reginaldalfret in https://github.com/NVIDIA/cuopt/pull/2068
• FIX : respecting project_initial_primal in mPDLP by @Bubullzz in https://github.com/NVIDIA/cuopt/pull/2050
• Fix rationalize coefficient bug and too conservative knapsack separation for fractional capacity by @chris-maes in https://github.com/NVIDIA/cuopt/pull/2083
• grpc switch worker creation from fork to posix spawn with exec by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2085
• Local patch for papilo by @mlubin in https://github.com/NVIDIA/cuopt/pull/2102
• Keep presolve enabled for MIP set-solution callbacks by @akifcorduk in https://github.com/NVIDIA/cuopt/pull/2098
• fix(packaging): make single-component installs consumable from CMake by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2094
• build.sh: find libcuopt build tree for the cuopt Python build without --install by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2120
• Downgrade PSLP to v0.1.0 by @rg20 in https://github.com/NVIDIA/cuopt/pull/2116
• Fix leak of stale duals in compute_dual_solution_from_basis by @mlubin in https://github.com/NVIDIA/cuopt/pull/2124
• Fix duplicate substitution records emitted in presolve probing by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/2126
• ci: skip the PuLP nightly step on Python < 3.12 by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2138
• fix(mip): ask every early heuristic to stop before waiting for any by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2139

📖 Documentation

• docs: enforce heading title case with Vale by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1668
• Show mamba commands directly instead of a conda-then-swap note by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1827
• docs(grpc): document the VRP gRPC client (RoutingClient) by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1838
• docs: make Java bindings discoverable, documented, and installable via the selector by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1999
• docs build server spec from proxy server by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2084
• cuopt server documentation update by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2115
• docs: Release Notes 26.10 and missing settings docs by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1972

🚀 New Features

• Upgrade cmake version to include the PDLP objective offset fix by @Kh4ster in https://github.com/NVIDIA/cuopt/pull/1695
• Expose solver statistics as scalar solution attributes by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1715
• Add a solver setting for controlling adaptive regularization in barrier by @rg20 in https://github.com/NVIDIA/cuopt/pull/1743
• Java bindings for LP, MIP and QP by @nvidiacbrissette in https://github.com/NVIDIA/cuopt/pull/1524
• move server LPData transform into a standalone function by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1848
• Extract LP create_data_model/create_solver into conversion.py by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1849
• extract routing server data creation routines into standalone module by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1850
• move server data compression handling into a standalone module by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1851
• extract result file support logic to be shared by proxy by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1869
• Move deprecated server modules into utils/deprecated. by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1874
• move cuopt server job_queue and worker runtime into deprecated utils by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1876
• routing: add distance breaks by @np96 in https://github.com/NVIDIA/cuopt/pull/1196
• Keep permanent utils from importing the legacy HTTP runtime. by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1879
• add cancel and status for grpc routing client by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1886
• add basic http proxy server that forwards to grpc by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1892
• Add solver caching to support re-solves for barrier QP by @Iroy30 in https://github.com/NVIDIA/cuopt/pull/1821
• proxy server add VRP request and solution over gRPC by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1917
• proxy server make the health check dependent on live grpc server by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1919
• proxy server add pdlp warmstart support to match legacy by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1918
• add cuopt/cuopt sync endpoint to the proxy server for backward compatability by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1933
• Expose distributed (multi-GPU) PDLP settings in the Java API by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1961
• Implement Curtis Reid Scaling by @rg20 in https://github.com/NVIDIA/cuopt/pull/1934
• proxy server implement incumbent_set_solutions by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1971
• Fix logging bug when asking for more GPUs than available by @Bubullzz in https://github.com/NVIDIA/cuopt/pull/1935
• change CI to test proxy+grpc instead of legacy http by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1993
• grpc support a map<string, string> for parameters in the server by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2010
• Add C API support for multiGPU PDLP by @Bubullzz in https://github.com/NVIDIA/cuopt/pull/1958
• grpc modify the client to use the new parameter map<string, string> by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2022
• Distributed curtis reid by @Bubullzz in https://github.com/NVIDIA/cuopt/pull/2004
• grpc add a gpu health check before each job or after idle by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2031
• standardize interface and remote solves on solver_settings_t by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2034
• grpc add standard grpc health problems and fix up helm chart by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2047
• grpc for the bundled client with new server handle only new parameter map by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2057
• proxy server add legacy openapi spec endpoint for parity by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2077
• Fix multi-GPU PDLP dispatch in the gRPC server worker by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1962
• Implement column reduction for PDLP for large problems by @rg20 in https://github.com/NVIDIA/cuopt/pull/2067

🛠️ Improvements

• ci: sync skill-card.md versions in update-version flow by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1669
• fix: add statuses: read permission to request-nvskills-ci workflow by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1681
• docs: fix Vale heading case errors in gRPC async client docs by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1682
• Symbol visibility controls and exports (remove-internal approach) by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1625
• ensure nightly builds always produce new packages by @jameslamb in https://github.com/NVIDIA/cuopt/pull/1661
• ci: update NVSkills CI request template and add require-nvskills-status workflow by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1690
• chore: trigger NVSkills CI for cuopt-multi-objective-exploration by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1673
• chore: trigger NVSkills CI for cuopt-numerical-optimization-formulation by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1675
• chore: trigger NVSkills CI for cuopt-numerical-optimization-api by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1674
• Add iterative refinement to the GPU ADAT solve path by @rg20 in https://github.com/NVIDIA/cuopt/pull/1680
• Update to rapids-logger 0.3 by @bdice in https://github.com/NVIDIA/cuopt/pull/1659
• ci: make pr-test-summary and NVSkills CI checks non-blocking by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1696
• chore: trigger NVSkills CI for cuopt-routing-api-python by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1676
• chore: trigger NVSkills CI for cuopt-user-rules by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1679
• Add a hyperparameter that controls bounding of free variables by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/1713
• Print total setup time by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/1714
• make 'vale' check optional locally by @jameslamb in https://github.com/NVIDIA/cuopt/pull/1708
• fix: exec the binary from console-script wrappers instead of spawning it by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1723
• Update cpp/benchmarks codeowners to cuopt-solver-codeowners by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1742
• Release pslp presolve memory by @hlinsen in https://github.com/NVIDIA/cuopt/pull/1735
• wheels: avoid vendoring libnvJitLink by @jameslamb in https://github.com/NVIDIA/cuopt/pull/1741
• Build cuopt as an abi3 wheel using the CPython Limited API by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1731
• Routing over gRPC: VRP server + compiled C++/Cython client by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1597
• test: isolate library wheel smoke test by @bdice in https://github.com/NVIDIA/cuopt/pull/1624
• Build conda python packages independently of the Python version by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1744
• Mod-2 cut separator for zero-half cuts by @akifcorduk in https://github.com/NVIDIA/cuopt/pull/1726
• Revert presolve tolerance by @akifcorduk in https://github.com/NVIDIA/cuopt/pull/1757
• Block Bounded Value Elimination presolving pass for binary problems by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1683
• Add a binary 2opt move fed from the probing cache into CPUFJ by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1738
• Reuse existing cuOptGetProblemIntAttribute to simplify code in c API by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/1769
• ci: stop nightly builds for release/26.08 by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1777
• fix: give each solver its own seed instead of a process-wide counter by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1717
• Root heuristics by @nguidotti in https://github.com/NVIDIA/cuopt/pull/1700
• RENS by @nguidotti in https://github.com/NVIDIA/cuopt/pull/1719
• chore: trigger NVSkills CI for cuopt-install by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1672
• Add Kitmaker Portal wheel publishing via Charon Ferry by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1763
• ci: report step timeouts explicitly, and skip the test that was hitting them by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1773
• Add a dual aggregation presolver for binary variables with a single lock by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/901
• B&B Halt by @nguidotti in https://github.com/NVIDIA/cuopt/pull/1760
• Resolve NVSkills Tier 1 security findings in cuopt-server-api-python by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1677
• Add parameters for tuning barrier primal and dual regularization by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/1807
• workaround for signal during grpc test framework teardown by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1691
• Make MIP root relaxation honor LP method by @hlinsen in https://github.com/NVIDIA/cuopt/pull/1829
• Python api performance improvement by @Iroy30 in https://github.com/NVIDIA/cuopt/pull/1615
• Add more strategies of init point computation in barrier by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/1577
• Remove unused cusolver/curand deps from libcuopt by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1830
• Reduce GPU/host synchronization overhead in barrier termination check by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/1808
• refactor(logger): give each component library its own logger by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1778
• Fix vendored PSLP writing a stray infeasible-message printf to native stdout by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1836
• refactor: split host-only members out of CUDA translation units by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1801
• Exhaustive coefficient reduction for binary rows by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1686
• Heuristic for arc-flow formulations by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1833
• RoutingClient: match Client's (host, port, tls=...) signature by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1840
• Tune Maven's own retry/concurrency to reduce Maven Central rate-limiting by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1835
• feat(c-api): add cuOptSetLogCallback by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1636
• Diving on cut passes by @nguidotti in https://github.com/NVIDIA/cuopt/pull/1812
• Replace process-wide seed_generator with per-worker RNGs in MIP heuristics by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1809
• Replaced cusparse wrappers with simple unique_ptrs and more RAII by @Bubullzz in https://github.com/NVIDIA/cuopt/pull/1342
• Set Papilo back to the main branch by @nguidotti in https://github.com/NVIDIA/cuopt/pull/1694
• Adopt CUDA stream compatibility accessors by @bdice in https://github.com/NVIDIA/cuopt/pull/1858
• Bump cuopt-java compile target from Java 11 to Java 17 by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1865
• add a smoke test to ensure server images start correctly by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1646
• refactor: keep the GPU warm-start path out of host-only translation units by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1803
• update variable type by @Iroy30 in https://github.com/NVIDIA/cuopt/pull/1783
• build: add the CUDA-free cuopt_client library by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1804
• Improve duplicate cut detection by @hlinsen in https://github.com/NVIDIA/cuopt/pull/1764
• refactor: build the routing gRPC arm into cuopt_client by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1884
• Improvement for qcqp canonicalization and symmetric check by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/1864
• Presolve: remove empty columns with a diagonal quadratic term by @rg20 in https://github.com/NVIDIA/cuopt/pull/1885
• simplify test for incumbent solution arrival during wait by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1898
• [CPUFJ PR 1] Improve time-to-first-incumbent, unify solution callback publication, scaffolding changes by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1880
• chore(skills): trigger benchmark regeneration for cuopt-server-api-python by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1895
• Don't abort the process when cudssSetThreadingLayer fails by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1897
• [CPUFJ PR2] Add a fast-path to the CPUFJ engine for binary problems with integer activities by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1881
• fix(test): stop DeleteQueuedJobPreventsRun/DeleteRunningJobCancelsWorker from inheriting worker respawn debt by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1877
• add a proxy server CPU path for routing conversion to DataModel by @tmckayus in https://github.com/NVIDIA/cuopt/pull/1910
• Speedup barrier second-order cone conversion in general path by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/1914
• test: skip incumbent-callback tests until [#1909] lands by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1921
• Fix compilation with Clang / LLVM by @vitor1001 in https://github.com/NVIDIA/cuopt/pull/1843
• refactor(routing): standardize cost terminology by @Juan-Francisco-Robles in https://github.com/NVIDIA/cuopt/pull/1299
• Describe solver settings and defaults in the gRPC codegen field registry by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1701
• build(cython): link each extension module against the component it uses by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1920
• Update Doxygen to 1.18.0 by @vuule in https://github.com/NVIDIA/cuopt/pull/1900
• Give cuDSS a threading layer built against cuOpt's own libgomp (wheels) by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1905
• Use new verify-dependencies hook from pre-commit-hooks by @KyleFromNVIDIA in https://github.com/NVIDIA/cuopt/pull/1915
• Handle PSLP false-infeasible on badly-scaled LPs by @rg20 in https://github.com/NVIDIA/cuopt/pull/1922
• Clarify factorization issue causes residual error by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/1866
• build(conda): split libcuopt into per-component packages by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1928
• cuopt_mcp: add an MCP server for LP/MILP solving over gRPC by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1916
• Explore self-contained Java classifier JARs with a statically linked libcuopt by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1818
• Enable all NVCC warnings as errors by @mlubin in https://github.com/NVIDIA/cuopt/pull/1948
• build(cmake): stop compiling routing/mathopt sources twice by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1960
• Scale the variables with variable bounds by @rg20 in https://github.com/NVIDIA/cuopt/pull/1652
• fix(client): hold the GPU variant alternatives behind an opaque pointer by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1940
• Fusing Halpern Kernels for performances by @Bubullzz in https://github.com/NVIDIA/cuopt/pull/1904
• Build the Java bindings dynamically against the pip-installed libcuopt in the official Docker images by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1959
• codeowners: route gRPC paths under cpp/ to infra team by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1975
• build(wheel): split libcuopt into per-component wheels by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1929
• Update PSLP to v0.0.12 by @hlinsen in https://github.com/NVIDIA/cuopt/pull/1943
• Add NNZ cutoff for concurrent mode by @hlinsen in https://github.com/NVIDIA/cuopt/pull/1932
• [CPUFJ PR3] Modular CPUFJ engine, simple portfolio by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1950
• cuopt-developer skill: sanitize internal context before it becomes public by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1988
• Primal and dual simplex improvements by @chris-maes in https://github.com/NVIDIA/cuopt/pull/1889
• Eliminate zero-cost free variables by equality substitution in QP/SOCP barrier presolve by @rg20 in https://github.com/NVIDIA/cuopt/pull/1912
• Fix O(n^2) loop in the flow cover cut generation by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1842
• build(wheel): drop the CUDA suffix and dimension from libcuopt-client by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1967
• fix(barrier): resolve cuDSS mtlayer path at runtime by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1982
• [CPUFJ PR4] Structure-aware starts and repairs for CPUFJ by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1951
• Rename CUOPT_DISTRIBUTED_PDLP_PARTITIONER to CUOPT_MULTIGPU_PDLP_PARTITIONER; remove CUOPT_USE_DISTRIBUTED_PDLP by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1984
• Include objective gap in QP termination criteria by @rg20 in https://github.com/NVIDIA/cuopt/pull/1733
• [Java] Remove setMethod/setPDLPSolverMode/setNumGpus/setMpdlpPartitioner wrappers by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2000
• [CPUFJ PR5] Portfolio diversification, more starts, move batching by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/1953
• Move Ruiz scaling and building KKT system to the GPU by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/1913
• Mutation by @nguidotti in https://github.com/NVIDIA/cuopt/pull/2001
• grpc integration test tighten up warmup job handling by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2012
• Indicator Strengthening by @nguidotti in https://github.com/NVIDIA/cuopt/pull/1983
• fixing mPDLP logs, adding header back by @Bubullzz in https://github.com/NVIDIA/cuopt/pull/2017
• perf(wheel): skip gRPC/abseil build for mathopt and routing wheels by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2035
• Replace CUB umbrella header with specific includes by @bdice in https://github.com/NVIDIA/cuopt/pull/2033
• Remove remaining rmm::cuda_stream_view usage by @bdice in https://github.com/NVIDIA/cuopt/pull/2030
• ci: gate wheel publish on the whole libcuopt/cuopt family building by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2051
• perf(wheel): skip mathopt/routing/grpc_server when building libcuopt-client by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2043
• Add update apis - RHS by @Iroy30 in https://github.com/NVIDIA/cuopt/pull/1941
• ci: remove issue auto-labeling, round-robin assignment, and inactivity reminder by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2053
• chore(release): add manual script to publish placeholder release images by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2038
• fix(conda): declare libraft-headers, and stop routing inheriting nccl by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2060
• ci: use proxy cache for pypi.anaconda.org, remove pip indices from dependencies.yaml by @jameslamb in https://github.com/NVIDIA/cuopt/pull/2064
• fix(conda): drop nccl and libnvjitlink from the client's run exports by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2070
• Flag to expose symbols by @nguidotti in https://github.com/NVIDIA/cuopt/pull/2074
• ci: bump C++ gtest step timeout to 90 minutes by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2087
• grpc follow up to [#2063] by @tmckayus in https://github.com/NVIDIA/cuopt/pull/2071
• Remove Papilo unconditional NDEBUG definition by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/2078
• Wire cuOpt Java bindings into the shared maven-publish workflow by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/1970
• Reduce PDLP memory use and avoid warm-start copies by @rg20 in https://github.com/NVIDIA/cuopt/pull/2065
• fix(ci): report crashed gtest binaries in pr-test-summary by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2086
• Update PSLP to v0.1.1 by @rg20 in https://github.com/NVIDIA/cuopt/pull/2092
• Fix bug [#2042] absence of logs on mPDLP error with C API and not using rhs when present by @Bubullzz in https://github.com/NVIDIA/cuopt/pull/2091
• Propagate primal simplex to APIs. Ensure consistency with dual simplex. by @chris-maes in https://github.com/NVIDIA/cuopt/pull/2096
• [CPUFJ PR6] CDCL heuristic start for CPUFJ by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/2049
• Persistent LNS ruin-repair search by @akifcorduk in https://github.com/NVIDIA/cuopt/pull/2058
• update code base to replace occurences of "distributed PDLP" with multi-GPU PDLP by @Bubullzz in https://github.com/NVIDIA/cuopt/pull/2076
• Reduce barrier refinement allocations and launch overhead by @rg20 in https://github.com/NVIDIA/cuopt/pull/2069
• Update APIs - SOCP support by @Iroy30 in https://github.com/NVIDIA/cuopt/pull/1979
• Exploit degeneracy in MIP by @chris-maes in https://github.com/NVIDIA/cuopt/pull/1685
• Fix nightly build: barrier_cache_t template argument in SOCP barrier test by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2110
• Map LP initial solutions through PSLP presolve for PDLP by @rg20 in https://github.com/NVIDIA/cuopt/pull/2024
• Fix excessive setups and low diversity in local search CPUFJ calls by @aliceb-nv in https://github.com/NVIDIA/cuopt/pull/2105
• Skip duplicate banner print in sequential solve by @yuwenchen95 in https://github.com/NVIDIA/cuopt/pull/2109
• add a cli toggle for PDLP hyper parameter project_initial_primal by @Bubullzz in https://github.com/NVIDIA/cuopt/pull/2089
• ci: use shared NVSkills team-request workflow to support fork PRs by @ramakrishnap-nv in https://github.com/NVIDIA/cuopt/pull/2100
• CPUFJ: no per-move objective rebuild while infeasible and add halt checks in the epigraph projection loop by @hlinsen in https://github.com/NVIDIA/cuopt/pull/2117
• Preserve recoverable barrier iterates with converged objective gaps by @rg20 in https://github.com/NVIDIA/cuopt/pull/2103
• add a C API test to ensure mPDLP works with cuOptCreateProblem from RHS by @Bubullzz in https://github.com/NVIDIA/cuopt/pull/2108
• Limit column reduction log to first activation by @chris-maes in https://github.com/NVIDIA/cuopt/pull/2125
• ci(thirdparty-cvxpy): test the cvxpy release when no C++ compiler is available by @nicoloangileri in https://github.com/NVIDIA/cuopt/pull/2101

New Contributors

• @nvidiacbrissette made their first contribution in https://github.com/NVIDIA/cuopt/pull/1524
• @vgong-nv made their first contribution in https://github.com/NVIDIA/cuopt/pull/1888
• @Juan-Francisco-Robles made their first contribution in https://github.com/NVIDIA/cuopt/pull/1299
• @ZedingZhang made their first contribution in https://github.com/NVIDIA/cuopt/pull/1925
• @vuule made their first contribution in https://github.com/NVIDIA/cuopt/pull/1900
• @neuriv made their first contribution in https://github.com/NVIDIA/cuopt/pull/1947
• @derluke made their first contribution in https://github.com/NVIDIA/cuopt/pull/1775
• @reginaldalfret made their first contribution in https://github.com/NVIDIA/cuopt/pull/2068
• @nicoloangileri made their first contribution in https://github.com/NVIDIA/cuopt/pull/2101

Full Changelog: https://github.com/NVIDIA/cuopt/compare/v26.10.00a...v26.10.00