Products: ABAQUS/Standard ABAQUS/Explicit
Two types of parallel processing modes are available in ABAQUS. The thread-based mode:
is available on shared memory architecture platforms;
is available on all supported systems; and
is not supported by parallel element operations in ABAQUS/Standard.
is available for both shared and distributed memory architecture platforms;
is not available for Windows platforms;
is not available with the direct linear equation solver and the Lanczos solver in ABAQUS/Standard; and
requires that the MPI components be installed.
Output the local installation notes for your system to learn about local multiprocessing capabilities (see “Execution procedure for obtaining information,” Section 3.2.1). See the ABAQUS Home Page under Support & Services
Technical SupportSystems Information for the current version of ABAQUS for complete information about parallel processing support on various platforms, including information about MPI requirements and availability.
Parallel execution of ABAQUS is implemented using two different communication schemes: threads and Message Passing Interface (MPI). The MPI-based implementation is the default on all platforms where it is supported. The main advantage of the MPI-based parallel processing mode is that it allows parallel execution on both shared and distributed memory systems; the thread-based mode enables parallel execution only on shared memory systems.
Threads are lightweight processes that can perform different tasks simultaneously within the same application. Threads can communicate relatively easily by sharing the same memory pool.
Set mp_mode=threads on the command line to use this mode. Alternatively, set mp_mode=THREADS in the environment file (see “Using the ABAQUS environment settings,” Section 3.4.1). No additional software is required to run the thread-based parallelization mode.
MPI-based parallel runs use multiple analysis processes that communicate with each other via the Message Passing Interface (MPI). Set mp_mode=mpi on the command line to use this mode. Alternatively, set mp_mode=MPI in the environment file. This parallel processing mode is the default where it is available.
The MPI-based parallel mode can be used for parallel execution on multiple processors of a single machine (shared memory systems) or on multiple machines (distributed memory systems). For parallel execution on multiple machines or hosts, the list of hosts is given with the mp_host_list parameter. The file system may be shared or local when multiple hosts are used for the run. By default, ABAQUS determines the type of file system, but you can use the mp_file_system environment variable to specify the available file system.
See Chapter 4, “Customizing the ABAQUS environment,” of the ABAQUS Installation and Licensing Guide, for a complete description of the environment variables related to MPI-based parallel runs.
The following ABAQUS/Standard features can be executed in parallel: the direct sparse solver, the iterative solver, the Lanczos eigensolver, and element operations. The analysis input preprocessing is not executed in parallel. For ABAQUS/Explicit all of the computations other than those involving the analysis input preprocessor and the packager can be executed in parallel. Each of the features that are available for parallel execution has certain limitations, which are documented in detail; see “Parallel execution in ABAQUS/Standard,” Section 7.19.2, and “Parallel execution in ABAQUS/Explicit,” Section 7.19.3.
Figure 7.19.1–1 summarizes the available parallel processing modes for various features.
Some of the features can be executed in parallel using both modes, while others can be executed in parallel with only one of the two modes. To simplify usage of the parallel features in ABAQUS, it is recommended that the MPI-based parallel mode be used on all supported platforms. With this choice ABAQUS runs all available parallel features with MPI-based parallelization and activates thread-based parallel implementations for cases where an equivalent MPI-based implementation does not exist.
