Abstract
Structural topology optimization is a method of finding optimum material distribution within a given design domain, which is subjected to various loading and support conditions. However, the large computation time is one of the major challenges in its implementation. This challenge gets escalated with the use of unstructured mesh. In this paper, a Solid Isotropic Material with Penalization (SIMP)-based implementation of topology optimization on the graphics processing unit (GPU) for a \( 3{\text{D}}\) unstructured mesh is presented. The finite element analysis is performed entirely on a GPU. The main implementation challenges are addressed by developing an efficient and optimized GPU \( kernel\). The performance of the implementation is analyzed over an example using three different mesh sizes. Results show almost \( 4 \times\) speedup over a standard CPU implementation.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call