Distributed, Parallel, and Cluster Computing
SimSDP, a Rapid Prototyping tool for Radio Astronomy: From NenuFAR Experiments to SKAO-Scale Simulation
Published on - SiPS 2025 - 38th IEEE Workshop on Signal Processing Systems
SimSDP is a rapid prototyping tool designed to provide early estimates of the computation times and energy consumption of SKAO's future SDP processor pipeline. This paper extends a previous proof-of-concept by: (1) exploring further parallelism through dataflow modeling of spectral parallelism with joint or distributed deconvolution, (2) simulating pipeline behavior on multi-core multi-node CPU architectures for multiscale data, (3) deploying parametric prototype pipelines using real-world data from the NenuFAR instrument. Experimental results show that SimSDP effectively supports design-space exploration by simulating three radio-interferometric imaging algorithms on HPC systems, with simulation uncertainty below 10%.