Build against cuOpt 26.10 nightly: multi-GPU PDLP, new options, iteration/node reporting

[0x17]

• Switch both workflows to the latest cuOpt nightly wheels (>=26.10.0a0) from the RAPIDS nightly index
  • needs --prerelease=allow and --index-strategy unsafe-best-match so RAPIDS packages come from the nightly index and CUDA libraries from pypi.nvidia.com
• Adapt link and bundles to cuOpt 26.10's modular component wheels (both workflows and build-link.sh)
  • libcuopt is now a thin metapackage; libcuopt.so is a linker script, not the engine. Build against libcuopt_mathopt/include plus libcuopt_client/include, link -lcuopt_mathopt and the separate client library, and bundle libcuopt_mathopt/lib64/libcuopt_mathopt.so + libcuopt_client/lib64/libcuopt_client.so (routing is not needed by the GAMS link).
  • Copy libgomp, libtbb, libtbbmalloc and libcudart from libcuopt_mathopt_cuXX.libs, rather than libcuopt_cuXX.libs. The client also uses bundled libgomp/libcares: include both libcuopt_client_cuXX.libs and libcuopt_client.libs, since the unsuffixed client wheel may own the installed libcuopt_client.so. libomp is gone.
  • Copy cuOpt's cuDSS threading layer from libcuopt_mathopt/lib64/libcudss_mtlayer_cuopt.so; do not bundle libcudss_mtlayer_gomp.so. Read build version/hash from the mathopt wheel in build-link.sh instead of using placeholder defaults.
  • libcuopt_client.so needs system libssl.so.3/libcrypto.so.3/libz.so.1 (not bundled).
  • Verified local CUDA 12 and CUDA 13 builds and their ELF dependencies; PR CI builds and packages both CUDA variants on x86_64 and ARM64 (no multi-GPU solve in CI).
• Support multi-GPU PDLP (Add C API support for multiGPU PDLP  NVIDIA/cuopt#1958)
  • optcuopt.def: num_gpus now accepts -1..72, new multigpu_pdlp_partitioner option
  • bug fix: gmscuopt.c creates the problem with cuOptCreateRangedProblem instead of cuOptCreateProblem, since cuOpt's multi-GPU path only reads constraint lower/upper bounds and saw 0 constraints (validation error with presolve 0)
  • skip initial primal/dual solutions for multi-GPU PDLP, cuOpt rejects them
  • README section on usage and limitations
• Update optcuopt.def to cuOpt 26.10
  • remove 4 options dropped upstream (mip_hyper_heuristic_presolve_time_ratio, mip_hyper_heuristic_presolve_max_time, mip_hyper_diving_min_node_depth, mip_hyper_submip_node_limit_base)
  • new method 4 (primal simplex) and barrier_dual_initial_point 2 (SeDuMi-style), fix changed default of mip_hyper_heuristic_related_vars_time_limit
  • add 18 new options (e.g. concurrent_nnz_cutoff, primal_simplex_pricing, mip_rens, mip_mutation, barrier regularization, Curtis-Reid scaling); sequence_solve left out (Python re-solve cache only)
• Report iterations and nodes to GAMS via the new solution attributes (cuOptGetSolutionIntAttribute)
  • LP iterations, MIP nodes and simplex iterations; node-limit detection now uses the actual node count
  • guarded by #ifdef, still compiles against 26.08 headers
• Pass GAMS nodlim to cuOpt node_limit (was ignored before)
• Bug fix: drop the PDLP reduced-cost workaround in gmscuopt.c, cuOpt now returns correct reduced costs (Correctly return reduced costs for PDLP (stable3) NVIDIA/cuopt#1797)
• Verified locally against nightly 26.10.0a222 (CUDA 13, single RTX A1000):
  • regression tests 10/10, gamslib test 73/73 match CPLEX
  • multi-GPU code path (method 1, num_gpus -1, runs even on 1 GPU): 33/35 gamslib LPs match CPLEX within 1e-3; egypt hits the iteration limit (also with single-GPU PDLP), indus89 doesn't converge within 120 s (single-GPU PDLP: optimal in 25 s)
  • iterations/nodes and nodlim checked on trnsport and cube
  • not tested on a machine with several GPUs
• Not done: cuOptSetLogCallback for a live log, since it only receives lines from the calling thread (a MIP log loses ~75% of its lines)

[mlubin]

Huh? please let us know about these types of issues :)
@Bubullzz

[0x17]

Ah, you're right. Didn't have the time to properly structure this finding. The issue is now here NVIDIA/cuopt#2042.