Improving Computational Performance through HPC Techniques: case study using DD3IMP in‐house code. The computational efficiency of the FEA is strongly dependent on the algorithmic and numerical efficiency of the FE solver. This is particularly important in case of implicit FE codes, such as DD3IMP, the in‐house static implicit FE solver under analysis in this work. This study describes the procedure adopted to identify the main computational bottlenecks of the FE solver in order to introduce the OpenMP directives and, consequently, to achieve a major speedup of the whole algorithm. The different parallelized branches of the code are tested using the well‐known square cup deep drawing example, considering different FE discretizations. The analysis of the preliminary results, concerning the CPU wall time, allows to demonstrate that the adoption of HPC techniques, such as the abovementioned OpenMP directives, enables to: (i) achieve a speedup factor close to the number of cores (in a single computer); (ii) solve a problem in a shorter time; (iii) solve a bigger problem in the same amount of time and, thus, (iv) achieve a better solution in a given amount of time.
Keywords for this software
References in zbMATH (referenced in 3 articles )
Showing results 1 to 3 of 3.
- Neto, D.M.; Oliveira, M.C.; Menezes, L.F.: Surface smoothing procedures in computational contact mechanics (2017)
- Neto, D.M.; Oliveira, M.C.; Menezes, L.F.; Alves, J.L.: Applying Nagata patches to smooth discretized surfaces used in 3D frictional contact problems (2014)
- Oliveira, M.C.; Alves, J.L.; Chaparro, B.M.; Menezes, L.F.: Study on the influence of work-hardening modeling in springback prediction (2007)