Presolve stopps giving feedback
AnsweredI have an Integer Model that I want to solve, and for a while I thought Gurobi is stuck in Presolve, because the feedback (coming around every 5 sec) stopped. I experimented, deactivating the presolve worked for me (and was quite fast, but there is a post about that).
However, I realized that the CPU and RAM are still used by Gurobi, so I decided to let the PC work for a long time, and Gurobi came to a solution.
It would be nice, if Gurobi would continue giving regular feedback (it started with ~5 sec interval).
Here is the full log output:
Gurobi Optimizer version 9.0.0 build v9.0.0rc2 (win64)
Optimize a model with 2651761 rows, 1643280 columns and 10124064 nonzeros
Model fingerprint: 0xd9894306
Variable types: 0 continuous, 1643280 integer (1643280 binary)
Coefficient statistics:
Matrix range [1e+00, 1e+04]
Objective range [2e03, 9e02]
Bounds range [1e+00, 1e+00]
RHS range [1e+00, 1e+04]
Found heuristic solution: objective 0.1018125
Presolve removed 540005 rows and 12152 columns (presolve time = 6s) ...
Presolve removed 543845 rows and 12152 columns (presolve time = 10s) ...
Presolve removed 543845 rows and 12152 columns (presolve time = 15s) ...
Presolve removed 543845 rows and 12152 columns (presolve time = 20s) ...
Presolve removed 543845 rows and 12152 columns (presolve time = 25s) ...
Presolve removed 543845 rows and 12152 columns (presolve time = 30s) ...
Presolve removed 543845 rows and 12152 columns (presolve time = 35s) ...
Presolve removed 551525 rows and 12152 columns (presolve time = 41s) ...
Presolve removed 551525 rows and 12152 columns (presolve time = 48s) ...
Presolve removed 551525 rows and 237082 columns (presolve time = 6461s) ...
Presolve removed 841334 rows and 237155 columns (presolve time = 6465s) ...
Presolve removed 841336 rows and 237157 columns (presolve time = 6471s) ...
Presolve removed 841336 rows and 237157 columns
Presolve time: 6471.28s
Presolved: 1810425 rows, 1406123 columns, 8172761 nonzeros
Variable types: 0 continuous, 1406123 integer (1406123 binary)
Deterministic concurrent LP optimizer: primal simplex, dual simplex, and barrier
Showing barrier log only...
Root barrier log...
Ordering time: 1.39s
Barrier performed 0 iterations in 6492.51 seconds
Barrier solve interrupted  model solved by another algorithm
Concurrent spin time: 1.75s
Solved with dual simplex
Root relaxation: objective 0.000000e+00, 15 iterations, 19.12 seconds
Nodes  Current Node  Objective Bounds  Work
Expl Unexpl  Obj Depth IntInf  Incumbent BestBd Gap  It/Node Time
* 0 0 0 0.0000000 0.00000 0.00%  6492s
Explored 0 nodes (15 simplex iterations) in 6493.54 seconds
Thread count was 8 (of 8 available processors)
Solution count 2: 0 0.101812
Optimal solution found (tolerance 1.00e04)
Best objective 0.000000000000e+00, best bound 0.000000000000e+00, gap 0.0000%"

Hi Dániel,
Apologies for the (very) late reply!
Is this problem still occurring with the latest version? You could send us the model file so we can have a look to find out what is taking so much time without reporting about the progress.
Most likely, a very expensive presolving routine is responsible for this. Since your model is not *too* large, such a long pause in the log is not expected, though.
Cheers,
Matthias0
Please sign in to leave a comment.
Comments
1 comment