Controlling job parallel execution

To locate the job editor, select JobEditjobname from the main menu bar. Use the Parallelization tabbed page to control the parallel execution of an Abaqus analysis job, as follows:

See Also
Understanding analysis jobs
Using the job editor
Creating, editing, and manipulating co-executions
In Other Guides
Parallel Execution
  • Select the number of processors to use for the analysis if parallel processing is available.

  • For an Abaqus/Standard analysis, you can choose to use GPGPU acceleration of the direct sparse solver and specify the number of GPGPUs. For more information, see Parallel Execution in Abaqus/Standard.

  • For an Abaqus/Explicit analysis, you can choose the following:

    • The number of domains. If you choose domain-level parallelization, the number of domains must be equal to or a multiple of the number of processors.

      When the number of domains is a multiple of the number of processors, you can toggle on Activate dynamic load balancing if desired. This option may improve the performance of your analysis job. If you have 4 processors and choose 8, 12, 16, etc. domains, for example, the dynamic load balancing feature becomes available. Dynamic load balancing is turned off by default. This option is equivalent to the dynamic_load_balancing option in the Abaqus/Explicit execution procedure. For more information, see Abaqus/Standard and Abaqus/Explicit Execution.

      Dynamic load balancing is most likely to improve the computational speed and efficiency in applications with a strongly time-dependent and/or spatially varying computational load. Most imbalanced problems will see an optimal performance improvement when the number of domains is two to four times the number of processors. For more details, see Domain-Level Parallelization.

    • Whether the parallelization method should be Domain (default) or Loop. The domain-level method splits the model into a number of topological domains that are distributed evenly among the available processors. The loop-level method parallelizes low-level loops in the code that are responsible for most of the computational cost.

      If you select Loop for the parallelization method, you must select Default for the multiprocessing mode. For more information, see About Parallel Execution.

    For more information, see Parallel Execution in Abaqus/Explicit.

  • For both an Abaqus/Standard and an Abaqus/Explicit analysis, you can choose whether the multiprocessing mode should be Default, Threads, MPI (Message Passing Interface), or Hybrid. The default multiprocessing mode is dependent on the platform on which the analysis product executes.

    • For the Hybrid multiprocessing mode, you can choose the number of threads per MPI process.

    • You cannot choose the number of threads per MPI process for all other multiprocessing modes. For MPI and Default, the number of threads per MPI process has value of 1. For Threads, the number of threads per MPI process is equal to the number of processors.

    Note:

    You use the Step module to choose between the Abaqus/Standard iterative and direct sparse solvers. For more information, see Using the step editor.