Assessing the evaluation transformer model of reduction on the spineless G-machine

This paper reports on our initial work on assessing how using the evaluation transformer model of reduction affects the performance of lazy functional programs. The model uses information about how much evaluation of an expression is required in order to evaluate an expression as much as possible, as early as possible. Our results show that there is a definite gain over just using strictness information, but that it is difficult to characterise exactly how much gain there will be, and in what programs it will occur.