> The overhead of searching for symmetries and the exposition of dynamic clauses, make it a difficult trade off against the ruthless efficiency of traditional CDCL
That is certainly true. However, once all the other components have been optimized the hell out of them (of course that is a never ending story itself but probably with diminishing returns) the optimization of the symmetry breaking parts will follow. As I mentioned before, efficiently dealing with isomorphism in the search space is the ultimate (theoretical) optimization.
This development can already been observed looking at the history of graph automorphism packages. For a very long time "naughty" was the only significant player than suddenly came "saucy" and "bliss". As far as I know, one of the motivations for the newcomers was to use them for symmetry breaking.
That is certainly true. However, once all the other components have been optimized the hell out of them (of course that is a never ending story itself but probably with diminishing returns) the optimization of the symmetry breaking parts will follow. As I mentioned before, efficiently dealing with isomorphism in the search space is the ultimate (theoretical) optimization.
This development can already been observed looking at the history of graph automorphism packages. For a very long time "naughty" was the only significant player than suddenly came "saucy" and "bliss". As far as I know, one of the motivations for the newcomers was to use them for symmetry breaking.