IBM ILOG CPLEX Optimizer performance benchmarks

IBM ILOG CPLEX Optimizer confirms its performance leadership. The Version 12.4 CPLEX Optimizers provide yet another breakthrough in performance.

Mixed integer programming models are solved on average 15% faster than the industry-leading prior release 12.3, on models requiring 1 second and above, and for the more challenging group of models requiring at least 1000 seconds the average speedup is 40%.

The Barrier optimizer for linear programming models offers, as well, an average 10% faster solution speed than with the prior release for models requiring 1 second and above and 60% average speedup for models requiring 100 seconds or more.

The Concurrent optimizer for linear programming models solves models requiring 1 second or more an average of 30% faster than the 12.3 release and solves models requiring 100 seconds or more an average of 50% faster.

Comparison between CPLEX 12.3 (released June 2011) with CPLEX 12.4 (released December 2011) on 12 threads

CPLEX 12.4 Mixed integer linear programming (MILP) enhancement

CPLEX 12.4 Concurrent LP Enhancement

CPLEX 12.4 Barrier LP Enhancement

Benchmark conditions

Benchmarks were conducted on 2.66 GHz Intel Xeon X5650 processors with 2 x 6 cores, 12 Mbytes cache on each processor, and 24 Gbytes of RAM. The version of Linux used was 2.6.32-71.14.1.el6.x86_64 #1 SMP x86_64 GNU/Linux. Default algorithmic settings within CPLEX were used, with a time limit of 10,000 seconds and a tree memory limit of 6 Gbytes and utilizing 12 threads. The test set consisted of optimization models collected from public and private sources. Models unable to be solved within the time limit by all the versions were excluded. When a version is unable to solve a model within the time limit, the solve time is considered to be the time limit. Models judged to have numerically unstable formulations are excluded. A model is assigned to a timing category by the longest time of any of the versions. The parallel mode for the results with CPLEX 11 and above is the default deterministic mode; with earlier versions the parallel mode is opportunistic. The comparisons are done with geometric means.

CPLEX has a long history of constant performance improvement. You can compare the CPLEX speedups over the past ten years.

• ILOG CPLEX 12.4 (2011): 15% overall, 1.4X on 1000 seconds and up

• ILOG CPLEX 12.3 (2011): 20% overall, 2X on 1000 seconds and up

• ILOG CPLEX 12.2 (2010): 50% overall, 2.7X on 1,000 seconds and up

• ILOG CPLEX 12.0 (2009): 30% overall, 2X on 1,000 seconds and up

• ILOG CPLEX 11 (2007): 15% under one minute, 3X on 1-60 minutes, 10X on one hour and up

• ILOG CPLEX 10 (2006): 35% overall, 70% on “particularly difficult models”

• ILOG CPLEX 9 (2003): 50% on “difficult customer models”

• ILOG CPLEX 8 (2002): 40% overall, 70% on “difficult problems”

• ILOG CPLEX 7 (2000): 60% on “hard mixed integer problems”