Ophélie RENAUD (SATIE — IRISA)

SimSDP: Proof of Concept for Radio Astronomy Imaging on High-Performance Architectures

Resume by AI: The presentation focuses on SimSDP, a proof-of-concept framework for simulating radio astronomy imaging pipelines on high-performance computing (HPC) architectures for the Square Kilometre Array (SKA) project. The key objectives include optimizing resource allocation, analyzing imaging pipelines, and ensuring automatic workload balancing across multi-node and multi-core systems. The work emphasizes the integration of radio astronomy imaging algorithms, parallelism (dataflow models, clustering), and the automation of performance benchmarking across various architectures (CPU, GPU, multi-node). Ongoing efforts aim to automate benchmarking, validate results on HPC platforms like Grid5000, and enhance SimSDP’s accuracy and reliability.