Type Feedback vs. Concrete Type Inference: A Comparison ofOptimization Techniques for Object-Oriented Languages

Two promising optimization techniques for object-oriented languages are typefeedback (dynamic) and concrete type inference (static). We directly comparethe two techniques, evaluating their effectiveness on a suite of 23 SELFprograms while keeping other factors constant.Our results show that both systems inline >95% of all sends and deliver similaroverall performance with one exception: SELF\'s automatic coercion of machineintegers to arbitrary-precision integers upon overflow confounds type inferenceand slows down arithmetic-intensive benchmarks.We discuss several other issues which, given the comparable run-timeperformance, may influence the choice between type feedback and typeinference.