I think we should create a helper function to avoid multiple statics. Something like

    template <typename T, typename ET = amrex_enum_traits<T>,
              std::enable_if_t<ET::value,int> = 0>
    std::vector<std::pair<std::string,T>> getEnumNameValuePairs_helper ()
    {
            // the function body of current getEnumNameValuePairs
    }

    template <typename T, typename ET = amrex_enum_traits<T>,
              std::enable_if_t<ET::value,int> = 0>
    std::vector<std::pair<std::string,T>> const& getEnumNameValuePairs ()
    {
        static auto r = getNumNameValuePairs_helper<T>();
        return r;
    }

Can we even create an optimized version of getEnumNameString that avoids creating a list of all (unused) names altogether? @WeiqunZhang

Depending the size, maybe using std::map or std::unordered_map is better than std::vector. But it's hard to avoid building the whole list.

@WeiqunZhang ready from my end

@dpgrote can you please test if this solves your performance issue on WarpX 1D runs? :)

We could also test alternative versions using std::map or std::unordered_map.

This seems to work, thanks! I did a test comparing the time spent in the getExtractedName in WarpX. The original version was taking 0.8 seconds. With this PR, it is 0.1 seconds. With the PR in WarpX, it is 0.07 seconds. Not sure why that would be slightly faster.

Probably because the WarpX PR uses std::map.