error using AMPL-NEOS

194 views
Skip to first unread message

Hossein Haghighat

unread,
Aug 9, 2021, 8:35:45 AM8/9/21
to AMPL Modeling Language
hi,
I am solving an MIP problem with AMPL-Gurobi/CPLEX. I use the following setting 

option gurobi_options " nodefilestart 2.5  ";
option cplex_options " nodefile 3 ";

After several hours, I get the message below:

 "ERROR: your job was terminated because it exceed the maximum allowed amount of disk space. "

can you please let me know how to fix this ?


AMPL Google Group

unread,
Aug 9, 2021, 9:28:42 PM8/9/21
to AMPL Modeling Language
Since this is an "error using AMPL-NEOS", the error message must mean that CPLEX or Gurobi tried to use more disk space than NEOS allows for one job. Since this is a limitation imposed by the NEOS Server, it cannot be changed directly by setting any AMPL or solver option.

To learn more, you could turn on logging (CPLEX mipdisplay=2 or Gurobi outlev=1) and set a time limit (CPLEX time=sss or Gurobi timelim=sss, where sss is a number of seconds) so that your job will finish normally rather than being terminated by NEOS. Perhaps you will see that after sss seconds, the solver has already reduced the gap to a reasonable percentage, or has already gone a long time without finding any better solutions.

You could also try reducing the amount of storage needed for the tree by using fewer parallel threads -- though this may well make the job run longer. It appears that NEOS uses 4 threads by default, but you can try, say, threads=2 or threads=1 (in either CPLEX or Gurobi).


--
Robert Fourer
am...@googlegroups.com
{#HS:1595645545-105861#}

Hossein Haghighat

unread,
Aug 11, 2021, 7:42:24 AM8/11/21
to AMPL Modeling Language
thank you Bob for your reply.  I cannot still solve the model on NEOS, after applying the settings. here is the output of my model, which is terminated due to excessive time . I appreciate  very much  your help, offer and support.

You are using the solver cplexamp.

%% COMMENTS %%%%%%%%%%%%%%%%%%%%%%%%%%%%


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Checking ampl.mod for cplex_options...
Checking ampl.com for cplex_options...
Executing AMPL.
processing data.
processing commands.
Executing on prod-exec-6.neos-server.org

 outer CCG literation 1

SOLVING OUTER MASTER PROBLEM


Presolve eliminates 0 constraints and 1180 variables.
Adjusted problem:
2161 variables:
        1850 binary variables
        311 linear variables
354 constraints, all linear; 4320 nonzeros
        354 inequality constraints
1 linear objective; 2161 nonzeros.

CPLEX 20.1.0.0: nodefile 3
mipbasis 0
threads 2
mipdisplay 2
mipinterval 5000
Found incumbent of value 0.000000 after 0.00 sec. (0.04 ticks)

Root node processing (before b&c):
  Real time             =    0.00 sec. (0.06 ticks)
Parallel b&c, 2 threads:
  Real time             =    0.00 sec. (0.00 ticks)
  Sync time (average)   =    0.00 sec.
  Wait time (average)   =    0.00 sec.
                          ------------
Total (root+branch&cut) =    0.00 sec. (0.06 ticks)
CPLEX 20.1.0.0: optimal integer solution; objective 0
0 MIP simplex iterations
0 branch-and-bound nodes
No basis.



Presolve eliminates 0 constraints and 2130 variables.
Adjusted problem:
12691 variables:
        2390 binary variables
        10301 linear variables
17405 constraints, all linear; 90036 nonzeros
        3680 equality constraints
        9145 inequality constraints
        4580 range constraints
1 linear objective; 2161 nonzeros.



Start Inner CCG literation for outer iteration 1

SOLVING INNER MASTER PROBLEM, literation 1


Presolve eliminates 0 constraints and 21981 variables.
Adjusted problem:
24671 variables:
        910 binary variables
        23761 linear variables
11831 constraints, all linear; 64901 nonzeros
        8190 equality constraints
        3641 inequality constraints
1 linear objective; 1 nonzero.

CPLEX 20.1.0.0: nodefile 3
mipbasis 0
threads 2
mipdisplay 2
mipinterval 5000
Retaining values of one MIP start for possible repair.
Found incumbent of value 0.000000 after 0.01 sec. (8.08 ticks)
MIP Presolve eliminated 5450 rows and 8950 columns.
MIP Presolve modified 3480 coefficients.
Reduced MIP has 4791 rows, 14131 columns, and 42851 nonzeros.
Reduced MIP has 910 binaries, 0 generals, 0 SOSs, and 0 indicators.
Probing fixed 0 vars, tightened 29108 bounds.
Probing time = 0.50 sec. (600.21 ticks)
Detecting symmetries...
MIP Presolve eliminated 10 rows and 30 columns.
MIP Presolve modified 250 coefficients.
Reduced MIP has 4601 rows, 13921 columns, and 42541 nonzeros.
Reduced MIP has 910 binaries, 0 generals, 0 SOSs, and 0 indicators.
Probing time = 0.00 sec. (0.66 ticks)
MIP emphasis: balance optimality and feasibility.
MIP search method: dynamic search.
Parallel mode: deterministic, using up to 2 threads.
Root relaxation solution time = 0.18 sec. (29.92 ticks)

        Nodes                                         Cuts/
   Node  Left     Objective  IInf  Best Integer    Best Bound    ItCnt     Gap

*     0+    0                            0.0000                           0.00%
*     0+    0                       1.73863e+10                           0.00%
      0     0        cutoff         1.73863e+10   1.73863e+10     3564    0.00%
      0     0        cutoff         1.73863e+10   1.73863e+10     3564    0.00%
Elapsed time = 1.71 sec. (2954.29 ticks, tree = 0.01 MB)

Root node processing (before b&c):
  Real time             =    1.71 sec. (2954.84 ticks)
Parallel b&c, 2 threads:
  Real time             =    0.01 sec. (2.55 ticks)
  Sync time (average)   =    0.00 sec.
  Wait time (average)   =    0.00 sec.
                          ------------
Total (root+branch&cut) =    1.71 sec. (2957.39 ticks)
CPLEX 20.1.0.0: optimal integer solution; objective 17386278010
3564 MIP simplex iterations
0 branch-and-bound nodes
Tried aggregator 2 times
No basis.



SOLVING INNER SUB-PROBLEM


Presolve eliminates 0 constraints and 50312 variables.
Adjusted problem:
7820 variables:
        540 binary variables
        7280 linear variables
17050 constraints, all linear; 37680 nonzeros
        3680 equality constraints
        8790 inequality constraints
        4580 range constraints
1 linear objective; 1760 nonzeros.

CPLEX 20.1.0.0: nodefile 3
mipbasis 0
threads 2
mipdisplay 2
mipinterval 5000
Retaining values of one MIP start for possible repair.
Found incumbent of value 5.6954532e+10 after 0.01 sec. (7.09 ticks)
MIP Presolve eliminated 14160 rows and 8110 columns.
Reduced MIP has 1830 rows, 3230 columns, and 8540 nonzeros.
Reduced MIP has 0 binaries, 0 generals, 0 SOSs, and 0 indicators.
MIP Presolve eliminated 1647 rows and 2907 columns.
Reduced MIP has 183 rows, 323 columns, and 854 nonzeros.
Reduced MIP has 0 binaries, 0 generals, 0 SOSs, and 0 indicators.
Detecting symmetries...
MIP Presolve eliminated 2 rows and 2 columns.
Reduced MIP has 181 rows, 321 columns, and 841 nonzeros.
Reduced MIP has 0 binaries, 0 generals, 0 SOSs, and 0 indicators.
MIP emphasis: balance optimality and feasibility.
MIP search method: dynamic search.
Parallel mode: deterministic, using up to 2 threads.
Root relaxation solution time = 0.00 sec. (1.80 ticks)

        Nodes                                         Cuts/
   Node  Left     Objective  IInf  Best Integer    Best Bound    ItCnt     Gap

*     0+    0                       1.39099e+10   8.42117e+09            39.46%
*     0     0      integral     0   8.86741e+09   8.86741e+09      133    0.00%
Elapsed time = 0.06 sec. (66.47 ticks, tree = 0.00 MB)

Root node processing (before b&c):
  Real time             =    0.06 sec. (66.57 ticks)
Parallel b&c, 2 threads:
  Real time             =    0.00 sec. (0.00 ticks)
  Sync time (average)   =    0.00 sec.
  Wait time (average)   =    0.00 sec.
                          ------------
Total (root+branch&cut) =    0.06 sec. (66.57 ticks)
CPLEX 20.1.0.0: optimal integer solution; objective 8867412558
133 MIP simplex iterations
0 branch-and-bound nodes
absmipgap = 9.53674e-06, relmipgap = 1.07548e-15
No basis.

Start Inner CCG literation for outer iteration 1

SOLVING INNER MASTER PROBLEM, literation 2


Presolve eliminates 0 constraints and 39461 variables.
Adjusted problem:
47521 variables:
        910 binary variables
        46611 linear variables
22752 constraints, all linear; 128522 nonzeros
        15470 equality constraints
        7282 inequality constraints
1 linear objective; 1 nonzero.

CPLEX 20.1.0.0: nodefile 3
mipbasis 0
threads 2
mipdisplay 2
mipinterval 5000
Retaining values of one MIP start for possible repair.
Found incumbent of value 0.000000 after 0.03 sec. (15.36 ticks)
MIP Presolve eliminated 9430 rows and 16310 columns.
MIP Presolve modified 6960 coefficients.
Reduced MIP has 10142 rows, 28031 columns, and 87602 nonzeros.
Reduced MIP has 910 binaries, 0 generals, 0 SOSs, and 0 indicators.
Probing fixed 0 vars, tightened 79255 bounds.
Probing time = 1.40 sec. (1524.12 ticks)
Detecting symmetries...
MIP Presolve eliminated 49 rows and 109 columns.
MIP Presolve modified 500 coefficients.
Reduced MIP has 9733 rows, 27562 columns, and 86855 nonzeros.
Reduced MIP has 910 binaries, 0 generals, 0 SOSs, and 0 indicators.
Probing time = 0.00 sec. (1.29 ticks)
MIP emphasis: balance optimality and feasibility.
MIP search method: dynamic search.
Parallel mode: deterministic, using up to 2 threads.
Root relaxation solution time = 1.87 sec. (1812.81 ticks)

        Nodes                                         Cuts/
   Node  Left     Objective  IInf  Best Integer    Best Bound    ItCnt     Gap

*     0+    0                            0.0000                           0.00%
      0     0   5.89060e+09   360        0.0000   5.89060e+09     9821     ---
      0     0   5.89060e+09   185        0.0000     Cuts: 161    10332     ---
      0     0   5.89060e+09   174        0.0000      Cuts: 17    10454     ---
Repair heuristic found nothing.
*     0+    0                       5.89060e+09   5.89060e+09             0.00%
      0     0        cutoff         5.89060e+09   5.89060e+09    10454    0.00%
Elapsed time = 10.53 sec. (12403.72 ticks, tree = 0.01 MB)

Flow cuts applied:  129
Flow path cuts applied:  22

Root node processing (before b&c):
  Real time             =   10.53 sec. (12404.76 ticks)
Parallel b&c, 2 threads:
  Real time             =    0.00 sec. (0.00 ticks)
  Sync time (average)   =    0.00 sec.
  Wait time (average)   =    0.00 sec.
                          ------------
Total (root+branch&cut) =   10.53 sec. (12404.76 ticks)
CPLEX 20.1.0.0: optimal (non-)integer solution; objective 5890602139
10454 MIP simplex iterations
0 branch-and-bound nodes
Tried aggregator 2 times
No basis.
7 integer variables rounded (maxerr = 6.52395e-07).
Assigning integrality = 3e-07 might help.
Currently integrality = 1e-05.

INFEASIBLE INNER MASTER PROBLEM

nCUT = 1
inner_iter = 0
GAP_o = 17386300000

:     gap_i      innL_down      innU_up   L_down      U_up        :=
1   8518870000   8867410000   17386300000    0     17386300000
;


 outer CCG literation 2

SOLVING OUTER MASTER PROBLEM


Presolve eliminates 0 constraints and 77001 variables.
Adjusted problem:
9981 variables:
        2390 binary variables
        7591 linear variables
17405 constraints, all linear; 86166 nonzeros
        3680 equality constraints
        9145 inequality constraints
        4580 range constraints
1 linear objective; 2161 nonzeros.

CPLEX 20.1.0.0: nodefile 3
mipbasis 0
threads 2
mipdisplay 2
mipinterval 5000
Retaining values of one MIP start for possible repair.
MIP Presolve eliminated 6098 rows and 5440 columns.
MIP Presolve modified 15500 coefficients.
Reduced MIP has 10537 rows, 8351 columns, and 77436 nonzeros.
Reduced MIP has 1850 binaries, 0 generals, 0 SOSs, and 0 indicators.
Found incumbent of value 3.7909036e+10 after 0.23 sec. (241.51 ticks)
Probing fixed 0 vars, tightened 286 bounds.
Probing time = 0.02 sec. (9.87 ticks)
Detecting symmetries...
MIP Presolve eliminated 20 rows and 20 columns.
MIP Presolve modified 200 coefficients.
Reduced MIP has 10517 rows, 8331 columns, and 77306 nonzeros.
Reduced MIP has 1850 binaries, 0 generals, 0 SOSs, and 0 indicators.
Probing time = 0.00 sec. (4.13 ticks)
Clique table members: 185.
MIP emphasis: balance optimality and feasibility.
MIP search method: dynamic search.
Parallel mode: deterministic, using up to 2 threads.
Root relaxation solution time = 1.74 sec. (1698.57 ticks)

        Nodes                                         Cuts/
   Node  Left     Objective  IInf  Best Integer    Best Bound    ItCnt     Gap

*     0+    0                       3.79090e+10        0.0000           100.00%
*     0+    0                       6.84296e+09        0.0000           100.00%
*     0+    0                       6.80426e+09        0.0000           100.00%
      0     0   6.36183e+09    42   6.80426e+09   6.36183e+09    11769    6.50%
*     0+    0                       6.77307e+09   6.36183e+09             6.07%
      0     0   6.37252e+09    51   6.77307e+09     Cuts: 337    13194    5.91%
      0     0   6.37519e+09    52   6.77307e+09     Cuts: 182    19563    5.87%
*     0+    0                       6.64826e+09   6.37519e+09             4.11%
      0     0   6.37596e+09    57   6.64826e+09     Cuts: 139    20383    4.10%
*     0+    0                       6.56870e+09   6.37596e+09             2.93%
      0     0   6.37649e+09    55   6.56870e+09      Cuts: 59    21336    2.93%
      0     0   6.37666e+09    58   6.56870e+09   MIRcuts: 34    25974    2.92%
      0     0   6.37670e+09    59   6.56870e+09   MIRcuts: 23    26033    2.92%
Repair heuristic found nothing.
      0     2   6.37670e+09    59   6.56870e+09   6.37670e+09    26033    2.92%
Elapsed time = 17.09 sec. (16735.88 ticks, tree = 0.02 MB)
*    40+   34                       6.45060e+09   6.37670e+09             1.15%
*    40+   34                       6.44638e+09   6.37670e+09             1.08%
*   150+  149                       6.43537e+09   6.37670e+09             0.91%
*   707+  559                       6.43374e+09   6.37670e+09             0.89%
*   722+  558                       6.43355e+09   6.37670e+09             0.88%
*   733+  557                       6.43308e+09   6.37670e+09             0.88%
*   741+  557                       6.43307e+09   6.37670e+09             0.88%
*   755+  557                       6.43291e+09   6.37670e+09             0.87%

Performing restart 1

Repeating presolve.
MIP Presolve modified 12070 coefficients.
Reduced MIP has 10395 rows, 8209 columns, and 75194 nonzeros.
Reduced MIP has 1850 binaries, 0 generals, 0 SOSs, and 0 indicators.
MIP Presolve modified 6 coefficients.
Reduced MIP has 10395 rows, 8209 columns, and 75194 nonzeros.
Reduced MIP has 1850 binaries, 0 generals, 0 SOSs, and 0 indicators.
Represolve time = 0.17 sec. (175.24 ticks)
    843     0   6.37867e+09    60   6.43291e+09     Cuts: 147   160966    0.84%
    843     0   6.37928e+09    59   6.43291e+09      Cuts: 68   161297    0.83%
    843     0   6.37966e+09    75   6.43291e+09     Cuts: 115   164810    0.83%
    843     0   6.38004e+09    70   6.43291e+09      Cuts: 55   168015    0.82%
    843     0   6.38007e+09    74   6.43291e+09      Cuts: 98   171108    0.82%
    843     0   6.38009e+09    74   6.43291e+09      Cuts: 31   171565    0.82%
    843     0   6.38010e+09    72   6.43291e+09   MIRcuts: 44   174810    0.82%
    843     0   6.38012e+09    73   6.43291e+09    MIRcuts: 4   175386    0.82%
    843     0   6.38013e+09    74   6.43291e+09      Cuts: 29   178058    0.82%
    843     0   6.38013e+09    72   6.43291e+09      Cuts: 40   178118    0.82%
    843     0   6.38022e+09    70   6.43291e+09      Cuts: 20   185719    0.82%
    843     0   6.38043e+09    64   6.43291e+09      Cuts: 46   190156    0.82%
    843     0   6.38048e+09    67   6.43291e+09      Cuts: 52   196686    0.82%
    843     0   6.38053e+09    66   6.43291e+09      Cuts: 43   202510    0.81%
    843     0   6.38055e+09    71   6.43291e+09      Cuts: 48   212091    0.81%
    843     0   6.38056e+09    65   6.43291e+09      Cuts: 62   218290    0.81%
*  4510  3581      integral     0   6.42595e+09   6.38110e+09   539256    0.70%
*  4520  3417      integral     0   6.42522e+09   6.38110e+09   539575    0.69%
*  4643+ 3547                       6.42403e+09   6.38116e+09             0.67%
*  4643+ 3547                       6.42302e+09   6.38116e+09             0.65%
*  4643+ 3547                       6.41907e+09   6.38116e+09             0.59%
*  4643+ 3547                       6.41895e+09   6.38116e+09             0.59%
*  4643+ 3547                       6.41887e+09   6.38116e+09             0.59%
*  4643+ 3547                       6.41877e+09   6.38116e+09             0.59%
   5000  3657   6.39254e+09    47   6.41877e+09   6.38118e+09   580282    0.59%
  10000  8463   6.40392e+09    38   6.41877e+09   6.38142e+09  1099731    0.58%
  15000 13683   6.40685e+09    51   6.41877e+09   6.38170e+09  1658030    0.58%
* 17095+15699                       6.41753e+09   6.38186e+09             0.56%
* 17095+15699                       6.41619e+09   6.38186e+09             0.54%
* 17095+15699                       6.41566e+09   6.38186e+09             0.53%
* 17095+15699                       6.41465e+09   6.38186e+09             0.51%
* 17095+15699                       6.41430e+09   6.38186e+09             0.51%
* 17095+15699                       6.41410e+09   6.38186e+09             0.50%
* 17195+15375                       6.41408e+09   6.38187e+09             0.50%
  20000 18029   6.40314e+09    33   6.41408e+09   6.38196e+09  2135178    0.50%
  25000 22945   6.39840e+09    43   6.41408e+09   6.38233e+09  2614593    0.50%
  30000 27695   6.40244e+09    39   6.41408e+09   6.38255e+09  3083742    0.49%

Performing restart 2

Repeating presolve.
MIP Presolve eliminated 4 rows and 0 columns.
MIP Presolve modified 259 coefficients.
Reduced MIP has 10391 rows, 8209 columns, and 88486 nonzeros.
Reduced MIP has 1850 binaries, 0 generals, 0 SOSs, and 0 indicators.
MIP Presolve modified 162 coefficients.
Reduced MIP has 10391 rows, 8209 columns, and 88486 nonzeros.
Reduced MIP has 1850 binaries, 0 generals, 0 SOSs, and 0 indicators.
Represolve time = 0.20 sec. (152.77 ticks)
  33327     0   6.38065e+09    71   6.41408e+09     Cuts: 102  3412577    0.49%
  33327     0   6.38066e+09    67   6.41408e+09      Cuts: 38  3412643    0.49%
  33327     0   6.38070e+09    65   6.41408e+09      Cuts: 12  3412718    0.49%
  33327     0   6.38083e+09    67   6.41408e+09      Cuts: 30  3413861    0.49%
  33327     0   6.38085e+09    68   6.41408e+09      Cuts: 57  3414343    0.49%
  33327     0   6.38086e+09    66   6.41408e+09      Cuts: 23  3414417    0.49%
  33327     0   6.38087e+09    66   6.41408e+09      Cuts: 19  3414442    0.49%
  35000  1526   6.40508e+09    51   6.41408e+09   6.38263e+09  3575777    0.49%
  40000  6341   6.40879e+09    37   6.41408e+09   6.38326e+09  4026380    0.48%

ERROR: Your job was terminated because it exceeded the maximum allotted time for a job.

Please refer to the NEOS Server FAQ (https://neos-guide.org/content/FAQ) for more information regarding job termination conditions and potential solutions.

AMPL Google Group

unread,
Aug 12, 2021, 5:57:17 PM8/12/21
to AMPL Modeling Language
When CPLEX reaches the NEOS time limit, you are still in iteration 2, and the MIP gap for the master problem is still 0.48%. Thus it is looking doubtful that you will be able to finish your run using CPLEX, within the resource limitations imposed by NEOS.

Did you have the same experience with Gurobi? Sometimes one solver is much faster than another for a particular application, but the only way to tell is to try them both.


--
Robert Fourer
am...@googlegroups.com
{#HS:1595645545-105861#}
On Wed, Aug 11, 2021 at 11:42 AM UTC, AMPL Modeling Language <am...@googlegroups.com> wrote:
thank you Bob for your reply. I cannot still solve the model on NEOS, after applying the settings. here is the output of my model, which is terminated due to excessive time . I appreciate very much your help, offer and support.

You are using the solver cplexamp.

%% COMMENTS %%%%%%%%%%%%%%%%%%%%%%%%%%%%

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Checking ampl.mod for cplex_options...
Checking http://ampl.com/ for cplex_options...

Executing AMPL.
processing data.
processing commands.


outer CCG literation 1

SOLVING OUTER MASTER PROBLEM

Presolve eliminates 0 constraints and 1180 variables.
Adjusted problem:
2161 variables:
1850 binary variables
311 linear variables
354 constraints, all linear; 4320 nonzeros
354 inequality constraints
1 linear objective; 2161 nonzeros.

CPLEX http://20.1.0.0/: nodefile 3

mipbasis 0
threads 2
mipdisplay 2
mipinterval 5000
Found incumbent of value 0.000000 after 0.00 sec. (0.04 ticks)

Root node processing (before b&c):
Real time = 0.00 sec. (0.06 ticks)
Parallel b&c, 2 threads:
Real time = 0.00 sec. (0.00 ticks)
Sync time (average) = 0.00 sec.
Wait time (average) = 0.00 sec.
------------
Total (root+branch&cut) = 0.00 sec. (0.06 ticks)
CPLEX http://20.1.0.0/: optimal integer solution; objective 0

0 MIP simplex iterations
0 branch-and-bound nodes
No basis.

Presolve eliminates 0 constraints and 2130 variables.
Adjusted problem:
12691 variables:
2390 binary variables
10301 linear variables
17405 constraints, all linear; 90036 nonzeros
3680 equality constraints
9145 inequality constraints
4580 range constraints
1 linear objective; 2161 nonzeros.

Start Inner CCG literation for outer iteration 1

SOLVING INNER MASTER PROBLEM, literation 1

Presolve eliminates 0 constraints and 21981 variables.
Adjusted problem:
24671 variables:
910 binary variables
23761 linear variables
11831 constraints, all linear; 64901 nonzeros
8190 equality constraints
3641 inequality constraints
1 linear objective; 1 nonzero.

CPLEX http://20.1.0.0/: nodefile 3
CPLEX http://20.1.0.0/: optimal integer solution; objective 17386278010

3564 MIP simplex iterations
0 branch-and-bound nodes
Tried aggregator 2 times
No basis.

SOLVING INNER SUB-PROBLEM

Presolve eliminates 0 constraints and 50312 variables.
Adjusted problem:
7820 variables:
540 binary variables
7280 linear variables
17050 constraints, all linear; 37680 nonzeros
3680 equality constraints
8790 inequality constraints
4580 range constraints
1 linear objective; 1760 nonzeros.

CPLEX http://20.1.0.0/: nodefile 3
CPLEX http://20.1.0.0/: optimal integer solution; objective 8867412558

133 MIP simplex iterations
0 branch-and-bound nodes
absmipgap = 9.53674e-06, relmipgap = 1.07548e-15
No basis.

Start Inner CCG literation for outer iteration 1

SOLVING INNER MASTER PROBLEM, literation 2

Presolve eliminates 0 constraints and 39461 variables.
Adjusted problem:
47521 variables:
910 binary variables
46611 linear variables
22752 constraints, all linear; 128522 nonzeros
15470 equality constraints
7282 inequality constraints
1 linear objective; 1 nonzero.

CPLEX http://20.1.0.0/: nodefile 3
CPLEX http://20.1.0.0/: optimal (non-)integer solution; objective 5890602139
CPLEX http://20.1.0.0/: nodefile 3
On Tue, Aug 10, 2021 at 1:28 AM UTC, AMPL Google Group <am...@googlegroups.com> wrote:
Since this is an "error using AMPL-NEOS", the error message must mean that CPLEX or Gurobi tried to use more disk space than NEOS allows for one job. Since this is a limitation imposed by the NEOS Server, it cannot be changed directly by setting any AMPL or solver option.

To learn more, you could turn on logging (CPLEX mipdisplay=2 or Gurobi outlev=1) and set a time limit (CPLEX time=sss or Gurobi timelim=sss, where sss is a number of seconds) so that your job will finish normally rather than being terminated by NEOS. Perhaps you will see that after sss seconds, the solver has already reduced the gap to a reasonable percentage, or has already gone a long time without finding any better solutions.

You could also try reducing the amount of storage needed for the tree by using fewer parallel threads -- though this may well make the job run longer. It appears that NEOS uses 4 threads by default, but you can try, say, threads=2 or threads=1 (in either CPLEX or Gurobi).


--
Robert Fourer
am...@googlegroups.com

Hossein Haghighat

unread,
Aug 15, 2021, 10:37:19 AM8/15/21
to AMPL Modeling Language
here is the output of Gurobi on Neos. I set nodefilestart = 0.5, but it needs more disk space. isn't it a bug of Gurobi?



You are using the solver gurobi_ampl.

%% COMMENTS %%%%%%%%%%%%%%%%%%%%%%%%%%%%


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Checking ampl.mod for gurobi_options...
Checking ampl.com for gurobi_options...

Executing AMPL.
processing data.
processing commands.


 outer CCG literation 1

SOLVING OUTER MASTER PROBLEM


Presolve eliminates 0 constraints and 1180 variables.
Adjusted problem:
2161 variables:
        1850 binary variables
        311 linear variables
354 constraints, all linear; 4320 nonzeros
        354 inequality constraints
1 linear objective; 2161 nonzeros.

Gurobi 9.1.1: nodefilestart 0.5
basis 0
method 0
threads=4
Gurobi 9.1.1: optimal solution; objective 0
No dual variables returned.




Presolve eliminates 0 constraints and 2130 variables.
Adjusted problem:
12691 variables:
        2390 binary variables
        10301 linear variables
17405 constraints, all linear; 91116 nonzeros

        3680 equality constraints
        9145 inequality constraints
        4580 range constraints
1 linear objective; 2161 nonzeros.



Start Inner CCG literation for outer iteration 1

SOLVING INNER MASTER PROBLEM, literation 1


Presolve eliminates 0 constraints and 21981 variables.
Adjusted problem:
24671 variables:
        910 binary variables
        23761 linear variables
11831 constraints, all linear; 64901 nonzeros
        8190 equality constraints
        3641 inequality constraints
1 linear objective; 1 nonzero.

Gurobi 9.1.1: nodefilestart 0.5
basis 0
method 0
threads=4
Gurobi 9.1.1: optimal solution; objective 5.992271474e+10
9622 simplex iterations
No dual variables returned.




SOLVING INNER SUB-PROBLEM


Presolve eliminates 0 constraints and 50312 variables.
Adjusted problem:
7820 variables:
        540 binary variables
        7280 linear variables
17050 constraints, all linear; 38220 nonzeros

        3680 equality constraints
        8790 inequality constraints
        4580 range constraints
1 linear objective; 2300 nonzeros.

Gurobi 9.1.1: nodefilestart 0.5
basis 0
method 0
threads=4
Gurobi 9.1.1: optimal solution; objective 8370328597
5726 simplex iterations
1 branch-and-cut nodes
No dual variables returned.
1 integer variable rounded to integer; maxerr = 3.96766e-07


Start Inner CCG literation for outer iteration 1

SOLVING INNER MASTER PROBLEM, literation 2


Presolve eliminates 0 constraints and 39461 variables.
Adjusted problem:
47521 variables:
        910 binary variables
        46611 linear variables
22752 constraints, all linear; 128758 nonzeros

        15470 equality constraints
        7282 inequality constraints
1 linear objective; 1 nonzero.

Gurobi 9.1.1: nodefilestart 0.5
basis 0
method 0
threads=4
Gurobi 9.1.1: optimal solution; objective 8369179882
49746 simplex iterations
548 branch-and-cut nodes
No dual variables returned.
4 integer variables rounded to integers; maxerr = 9.20643e-06



SOLVING INNER SUB-PROBLEM


Presolve eliminates 0 constraints and 79162 variables.

Adjusted problem:
7820 variables:
        540 binary variables
        7280 linear variables
17050 constraints, all linear; 38220 nonzeros

        3680 equality constraints
        8790 inequality constraints
        4580 range constraints
1 linear objective; 2300 nonzeros.

Gurobi 9.1.1: nodefilestart 0.5
basis 0
method 0
threads=4
Gurobi 9.1.1: optimal solution; objective 8369268863
6364 simplex iterations
259 branch-and-cut nodes
No dual variables returned.
: innalpha_down   innL_down      innU_up      :=
1   8370330000    8370330000   59922700000
2   8369270000    8369270000    8369180000
;



lower BOUND of INNER CCG loop GREATER THAN upper


nCUT = 1
inner_iter = 0
GAP_o = 8369180000

:    gap_i    innL_down      innU_up   L_down      U_up       :=
1   88980.6   8370330000   59922700000    0     8369180000
2      .      8369270000    8369180000     .         .

;


 outer CCG literation 2

SOLVING OUTER MASTER PROBLEM


Presolve eliminates 0 constraints and 77001 variables.
Adjusted problem:
9981 variables:
        2390 binary variables
        7591 linear variables
17405 constraints, all linear; 87246 nonzeros

        3680 equality constraints
        9145 inequality constraints
        4580 range constraints
1 linear objective; 2161 nonzeros.

Gurobi 9.1.1: nodefilestart 0.5
basis 0
method 0
threads=4


ERROR: your job was terminated because it exceed the maximum allowed amount of disk space.

Please refer to the NEOS Server FAQ (https://neos-guide.org/content/FAQ) for more information regarding job termination conditions and potential solutions.


AMPL Google Group

unread,
Aug 16, 2021, 12:20:55 PM8/16/21
to AMPL Modeling Language
nodefilestart=0.5 tells Gurobi to use at most 0.5 gigabytes of computer memory for the nodes of the search tree. If the tree gets bigger than 0.5 gigabytes, then part of it is written out to disk. There is no limit to the size of the disk file that Gurobi might write, and so it is possible that the disk file could exceed NEOS's limit.

Adding outlev=1 will get you more information about how many nodes were in the search tree (and how fast the number of nodes was increasing) when NEOS terminated the run. Also it will show the size of the relative optimality gap at that point.

The NEOS Server FAQ mentions an 8 hour time limit and a 3 hour memory limit, but I can't see where it describes the disk space limit. You may need to write to sup...@neos-server.org to get information about allowed disk space.


--
Robert Fourer
am...@googlegroups.com
{#HS:1595645545-105861#}
On Sun, Aug 15, 2021 at 2:37 PM UTC, AMPL Modeling Language <am...@googlegroups.com> wrote:
here is the output of Gurobi on Neos. I set nodefilestart = 0.5, but it needs more disk space. isn't it a bug of Gurobi?

You are using the solver gurobi_ampl.

%% COMMENTS %%%%%%%%%%%%%%%%%%%%%%%%%%%%

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Checking ampl.mod for gurobi_options...
Checking http://ampl.com/ for gurobi_options...

Executing AMPL.
processing data.
processing commands.
On Thu, Aug 12, 2021 at 9:56 PM UTC, AMPL Google Group <am...@googlegroups.com> wrote:
When CPLEX reaches the NEOS time limit, you are still in iteration 2, and the MIP gap for the master problem is still 0.48%. Thus it is looking doubtful that you will be able to finish your run using CPLEX, within the resource limitations imposed by NEOS.

Did you have the same experience with Gurobi? Sometimes one solver is much faster than another for a particular application, but the only way to tell is to try them both.


--
Robert Fourer
am...@googlegroups.com
Reply all
Reply to author
Forward
0 new messages