12.07.2015 Views

ILOG CPLEX 11.0 User's Manual

ILOG CPLEX 11.0 User's Manual

ILOG CPLEX 11.0 User's Manual

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

the parallel MIP optimizer in deterministic mode with fewer than the maximum numberof available threads, set both the parallel mode and the threads parameter in yourapplication. See Determinism of Results on page 493 for background about this step.4. Call the parallel optimizer with the appropriate method or routine from Table 34.1.Table 34.1 Parallel Optimizer Methods and Routines of Component LibrariesOptimizer Concert IloCplex Method Callable LibraryParallel MIP Optimizer solve CPXmipoptParallel Barrier OptimizerConcurrent OptimizersetParam(RootAlg, Barrier)and then solvesetParam(RootAlg, Concurrent)and then solveCPXbaropt or CPXhybbaroptCPXsetintparam(env,CPX_PARAM_LPMETHOD,CPX_ALG_CONCURRENT)and then CPXlpopt or CPXqpoptParallel Barrier OptimizerThe <strong>ILOG</strong> <strong>CPLEX</strong> Parallel Barrier Optimizer achieves significant speedups over its serialcounterpart on a wide variety of classes of problems. (The serial Barrier Optimizer isintroduced in Solving LPs: Barrier Optimizer on page 197, and explored further in SolvingProblems with a Quadratic Objective (QP) on page 227 and in Solving Problems withQuadratic Constraints (QCP) on page 239.) Consequently, the parallel barrier optimizer willbe the best continuous choice on a parallel computer more frequently than on asingle-processor. For that reason, you should be careful not to apply performance data orexperience based on serial optimizers when you are choosing which optimizer to use on aparallel platform.Concurrent OptimizerOn a multiprocessor computer, the concurrent optimizer launches distinct LP and QPoptimizers on multiple threads, terminating as soon as the first optimizer finishes. The firstthread uses the same strategy as the single-processor automatic LPMethod setting (0). Ifa second thread is available, the concurrent optimizer runs the barrier optimizer on it. If athird processor is available, dual simplex, primal simplex, and barrier are all run. All furtheravailable threads are devoted to making the barrier optimization parallel. It should be notedthat the barrier optimization is not considered complete until the crossover step has beenperformed and simplex re-optimization has converged; in other words, regardless of which496 <strong>ILOG</strong> <strong>CPLEX</strong> <strong>11.0</strong> — USER’ S MANUAL

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!