White Papers

Abaqus Performance
17 Dell EMC Ready Solution for HPC Digital Manufacturing—Dassault Systѐmes’ Simulia Abaqus Performance
These results demonstrate that while the performance increase is not linear with respect to the number of
cores per node, there can be a substantial benefit in using system with large numbers of cores, and typically
the best performance is obtained when using all of the cores available. The exception for these cases is the
modal analysis model S3D, which is only thread parallel. The MPI mode is required to take full advantage of
several cores per server.
Figure 8 shows the parallel performance improvement when running a single job in parallel across multiple
nodes. These benchmarks were carried out on a cluster with four 6252 based servers using all available
cores in each server.
The parallel speedup when running jobs across more than a single node is somewhat mixed. These datasets
are rather small by current production standards, and do not represent typical production sized simulations.
However, these models do show significant speedup across two nodes, and modest speedups up to four
nodes.
Many FEA applications such as Abaqus, have been modified to allow for GPU acceleration with appropriate
GPU enabled servers. Figure 9 contains performance information for the Abaqus standard benchmarks
models described above run on a Dell EMC PowerEdge C4140 server, equipped with dual Intel Xeon Gold
20-core 6148 processors and four NVIDIA V100 GPUs.
1
1.25
1.5
1.75
2
2.25
2.5
1-node 2-node 4-node
Performance (relative to a single node)
Figure 8: Multi Server Parallel Performance
S4B S4D S6 E1 E3 E5