Comment by gshayban

The self-rewriting interpreter model is so so interesting, and has already proven itself in things like peak performance, and cross-language. However, it can have the disadvantage of blowing up warmup time. Are there Truffle language impls that persist their specialized interpreter nodes across executions?