Benchmark Ceridwen across GPU architectures

Benchmark Ceridwen systematically across different GPU architectures and hardware configurations to determine whether its reported sampling performance generalizes beyond the NVIDIA A100 and consumer-grade GPU tests described in the paper.

Background

The paper evaluates Ceridwen primarily on a single NVIDIA A100 GPU and reports that an initial test on an NVIDIA GeForce RTX 5060 achieved a comparable wall time. However, the authors do not provide a systematic comparison across GPU types, so the generality of the reported performance and speed-ups across hardware remains unresolved.

References

First tests on a consumer-grade GPU (an NVIDIA GeForce RTX 5060) reached the same wall time as the A100 for these models, but a systematic benchmark across GPUs is still outstanding.

— CERIDWEN: Fast and Flexible GPU-Accelerated Stellar Population Inference  (2609.30145 - Stoffers et al., 24 Sep 2026) in Section 5, Computational performance, Section 5.1