FWIW, I /believe/ that the Unified Model is bit-reproducible across different parallel decompositions (I'm not sure about >different compilations of the same code), and that's a very non-trivial code.
Yes, but they take a performance hit (because it limits the parallel preconditioners that they can use), and a large maintenance hit (some of the performance team seem to be entirely focussed on maintaining bit-reproducibility).
One possible scenario where you might wish for bit-reproducibility in an actual simulation run is computing adjoints of >chaotic systems with large Lyapunov exponents, where you really would like your replayed forward model to be >deterministic. However, I am by no means an expert, Patrick (if he's lurking) may have other opinions.
No, bit-reproducibility doesn't save you from numerical losses when differentiating chaotic maps. --cjc