Sort by
Refine Your Search
-
, improve, and use high-performance simulation tools for dynamic rupture, including GPU-oriented numerical methods You will verify and validate the numerical methods against benchmark problems and
-
for reproducible, efficient and scalable training and inference on parallel, distributed and GPU-accelerated computing systems Benchmark the developed approaches against established methods, assessing
-
realizations of the proposed latent reasoning algorithms. Quantitative comparisons with different hardware backends. Design conventional (GPU-based) deep neural networks for comparison. Publish research
-
) or HPC programming (MPI, OpenMP, CUDA, GPU computing) is considered an advantage but is not required Motivation to conduct excellent scientific research and publish in leading international journals
-
collaborate with academic partners in Europe and Asia, including research visits abroad. In this project, extensive collaboration is planned with researchers from TU Delft, Yonsei University, and GIST. The
-
-level circuit design and simulation What We Offer We provide GPU computing for machine learning on Aalto's Triton high-performance computing cluster, alongside industry-standard circuit simulation tools
-
and European supercomputer facilities LUMI and Computerome, along with local small-scale GPU resources at the Niels Bohr Institute. Local SAXS instrument are available at both institutes, and managed