User Tools

Site Tools


solver:fatal_error_insufficient_memory_for_setup_of_optimization_vectors

** Fatal Error ** Insufficient memory for setup of optimization vectors

The message indicates that CONOPT ran out of memory. A similar error message triggered by the same problem is:

 ** Fatal Error **  Insufficient memory to continue the
                    optimization.
 
                    You must request more memory.

CONOPT3 calculates an amount of memory with some rounding up and then asks for this amount. Due to the way memory is organized it is not possible to ask for more later on so if for some reason it runs out of memory you get an error message. You should find more information in the CONOPT log which will have been copied to the lst file right after the Solve Summary. That information will include what the estimate was that was insufficient. The current CONOPT system (23.7) is roughly limited to 1.6 GB on 32bit Windows, 3.8 GB on 32bit Linux and 8 GB on 64bit systems.

  • You can use the GAMS parameter workfactor to increase the multiplier for the allocation of solver memory: model_name.workfactor=1.5;. Here workfactor is set to 1.5 and this is the multiplier for CONOPT's estimation of memory. This can be used for models where CONOPT underestimates the memory requirement for the Hessian (and other parts of the model).
  • You can also specify a larger value for the memory requirements of CONOPT with model_name.workspace = xxx; before the Solve statement, where xxx is in megabytes. workspace is usually not a good choice if the model can vary a lot based on data. Again, 8000 MB is the maximum CONOPT can currently handle, if the model does not fit into this, you will not be able to solve this model with CONOPT.
solver/fatal_error_insufficient_memory_for_setup_of_optimization_vectors.txt · Last modified: 2011/08/03 13:00 by support