This report presents the significant advancements made in the UppASD software package, a tool for atomic spin dynamics simulations, through the integration of GPU acceleration. Originally developed at Uppsala University, UppASD primarily utilized CPU-based computations. This project aimed to enhance its computational efficiency and capability by extending and rigorously evaluating its GPU-based implementation. The focus was on integrating a variety of magnetic interactions, including tensorial coupling, anisotropy models, and the Dzyaloshinskii-Moriya interaction, into the GPU framework. We also assessed the impact of GPU acceleration on the calculation of key observables. The methodology involved a comprehensive evaluation of the existing code structure, meticulous debugging, and performance benchmarking. This report details the development process, highlights the challenges encountered, and provides insights into future directions for the UppASD project. / <p>Held internally sharing the progress made.</p>
| Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:uu-528136 |
| Date | January 2023 |
| Creators | Rieger, Philipp |
| Publisher | Uppsala universitet, Materialteori |
| Source Sets | DiVA Archive at Upsalla University |
| Language | English |
| Detected Language | English |
| Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess |
| Relation | FYSAST ; FYASPROJ1345 |
Page generated in 0.002 seconds