Decision Optimization

Decision Optimization

Delivers prescriptive analytics capabilities and decision intelligence to improve decision-making.


#Analytics
#DecisionOptimization
#DecisionOptimization
 View Only
  • 1.  CPXSwriteprob unaddressable access

    Posted 06/25/16 05:05 AM

    Originally posted by: 8699oriel


    Hi,

    I'm solving a QCP multiple times calling cplex via the C callable library.  In doing my memory analysis/cleanup with Dr. Memory (on Windows), every time I call CPXXwriteprob, Dr. Memory is reporting an "UNADDRESSABLE ACCESS beyond heap bounds: reading 8 byte(s)" in CPXEreadcopymipstarts.  The call chain is CPXXwriteprob : CPXSwriteprob : CPXSwriteparam : CPXEreadcopymipstarts : CPXEreadcopymipstarts.

     

    My primary question: should I be concerned about this, or can I ignore it?  

    My secondary question: why would a continuous problem (QCP) call a readcopymipstarts?

     

    Thank you,

    William


    #CPLEXOptimizers
    #DecisionOptimization


  • 2.  Re: CPXSwriteprob unaddressable access

    Posted 07/07/16 08:31 PM

    Are you able to narrow this down to a simple example and share it here?  It would be easier to provide guidance if we can reproduce the issue.  Just curious, which file format are you writing to (e.g., LP, MPS, SAV)?

     


    #CPLEXOptimizers
    #DecisionOptimization


  • 3.  Re: CPXSwriteprob unaddressable access

    Posted 07/13/16 01:25 AM

    Yes, you should be concerned about this. This is not expected.

    The callchain looks odd though. Do I understand correctly that it means that CPXSwriteprob() calls CPXSwriteparam() which in turn calls CPXEreadcopymipstarts()? If this is the claim then I think Dr. Memory is confused. I checked the source code and as expected, CPXSwriteprob() does not call CPXSwriteparam() and the latter does not call CPXEreadcopymipstarts().

    Like Ryan said, a small example to reproduce to issue would help.


    #CPLEXOptimizers
    #DecisionOptimization