[Cbc] Options not playing well together: -threads versus -extraVariables

acw at ascent.com acw at ascent.com
Mon Apr 22 10:35:11 EDT 2013


We are seeing a segfault from Cbc when both multi-threading and equal-cost 
variable aggregation are enabled.  The segfault occurs right after Cbc 
works through its root cuts.  The Cbc version is the trunk build of March 
20; it's not stable because Dr. Forrest had just added equal-cost variable 
aggregation.

Here is a log of Cbc output.  I can supply an MPS model, but it's the same 
one I sent to Dr. Forrest in March.

bin/cbc ../last-problem.mps -extrav 8 -threads 104 -solve 
Welcome to the CBC MILP Solver 
Version: Trunk (unstable) 
Build Date: Mar 20 2013 
Revision Number: 1876 

command line - bin/cbc ../last-problem.mps -extrav 8 -threads 104 -solve 
(default strategy 1) 
At line 1 NAME          BLANK 
At line 2 ROWS 
At line 1694 COLUMNS 
At line 7264 RHS 
At line 7664 BOUNDS 
At line 7883 ENDATA 
Problem BLANK has 1690 rows, 218 columns and 10834 elements 
Coin0008I BLANK read with 0 errors 
extraVariables was changed from 0 to 8 
threads was changed from 0 to 104 
Continuous objective value is 496912 - 0.01 seconds 
Cgl0003I 0 fixed, 35 tightened bounds, 2 strengthened rows, 0 
substitutions 
Cgl0004I processed model has 346 rows, 218 columns (218 integer) and 2983 
elements 
Problem has 218 integers (0 of which binary) and 0 continuous 
218 integer have nonzero objective, 0 have zero objective, 8 different 
nonzero (taking abs) 
48 variables have objective of 1 
1 variables have objective of 4 
28 variables have objective of 16.29 
28 variables have objective of 31.58 
8 variables have objective of 233.333 
35 variables have objective of 1235.43 
35 variables have objective of 1430.12 
35 variables have objective of 2441.3 
Replacing model - 7 new variables 
Cbc0012I Integer solution of 594089.87 found by DiveCoefficient after 0 
iterations and 0 nodes (0.19 seconds) 
Cbc0038I Full problem 353 rows 225 columns, reduced to 307 rows 135 
columns - too large 
Cbc0031I 11 added rows had average density of 185.27273 
Cbc0013I At root node, 11 cuts changed objective from 497140.24 to 
499066.93 in 100 passes 
Cbc0014I Cut generator 0 (Probing) - 0 row cuts average 0.0 elements, 4 
column cuts (4 active)  in 0.768 seconds - new frequency is 1 
Cbc0014I Cut generator 1 (Gomory) - 1621 row cuts average 213.4 elements, 
0 column cuts (11 active)  in 0.561 seconds - new frequency is 1 
Cbc0014I Cut generator 2 (Knapsack) - 0 row cuts average 0.0 elements, 0 
column cuts (0 active)  in 0.045 seconds - new frequency is -100 
Cbc0014I Cut generator 3 (Clique) - 0 row cuts average 0.0 elements, 0 
column cuts (0 active)  in 0.001 seconds - new frequency is -100 
Cbc0014I Cut generator 4 (MixedIntegerRounding2) - 0 row cuts average 0.0 
elements, 0 column cuts (0 active)  in 0.106 seconds - new frequency is 
-100 
Cbc0014I Cut generator 5 (FlowCover) - 0 row cuts average 0.0 elements, 0 
column cuts (0 active)  in 0.235 seconds - new frequency is -100 
Cbc0014I Cut generator 6 (TwoMirCuts) - 225 row cuts average 165.6 
elements, 0 column cuts (0 active)  in 0.143 seconds - new frequency is 
-100 
Segmentation fault (core dumped)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/cbc/attachments/20130422/9750b7a0/attachment.html>


More information about the Cbc mailing list