By Enzo Chen, Senior Engineer at Product R&D Division, Moldex3D
Under the trend of artificial intelligence, the demand for handling big data is emerging, driving the development of High-Performance Computing (HPC) platforms and devices. In the field of molding analysis, the utilization of HPC platforms can be widely applied in the development stages of plastics. The limitation of software specifications will no longer exist.
Nevertheless, as the products and manufacturing processes are getting complicated, great mesh amounts and accurate analysis require longer computing time. Repetitive virtual simulation also leads to a long development cycle. In order to speed up the mold filling analysis and successfully manufacture the products in the scheduled time, users need to choose between the computing efficiency and accuracy.
In Moldex3D 2020, the latest version, the solver has significantly enhanced the computing efficiency so that users can obtain the analysis results in a shorter time under the same hardware specification. The following case shows the comparison of Moldex3D R17 and 2020 in running filling analysis with 5, 12, and 20 million mesh elements. In Figure 1, with the CPU of AMD EPYC 7302 Processor, Moldex3D can help reduce 33%, 29%, and 20% calculation time respectively under the 8-core, 16-core, and 32-core computing.
The 10-core CPU used in Figure 2 is Intel Core i9-9900X CPU. In the calculation structure, 4 Intel CPUs are connected through a computer cluster. Under the 8-core, 16-core, and 32-core computing, Moldex3D 2020 can help reduce 50%, 27%, and 15% calculation time on average. Since the cluster calculation is limited by the internet transmission speed, the reduced computing time of 16-core and 32-core CPU is relatively less.
The optimization of the solver in Moldex3D 2020 can help reduce 30% calculation time on average with the same computing environment, mesh element amount, and process conditions. It enables users to enhance molding simulation efficiency and accelerate the product development process within a limited amount of time.
