The Definitive ANTLR 4 Reference, 2nd Edition

Maximizing Parser Speed

ANTLR v4’s adaptive parsing strategy is more powerful than v3’s, but it comes at the cost of a little bit of speed. If you need the most speed and the smallest memory footprint possible, you can do a two-step parsing strategy. The first step uses a slightly weaker parsing strategy, SLL(*), that almost always works. (It’s very similar to v3’s strategy, except it doesn’t need to backtrack.) If the first parsing step fails, you have to try the full LL(*) parse. After failing the first step, we don’t know whether it’s a true syntax error or whether it’s because the SLL(*) strategy wasn’t strong enough. Input that passes the SLL(*) step is guaranteed to pass the full LL(*), so there’s no point in trying out that more expensive ...