On Jan 7, 2010, at 10:04 AM, Marco Maggi wrote: > I guess that Ikarus itself is not to blame, rather it is > libffi which takes time.
Can you explain how you came to that conclusion? Yes, using libffi is definitely slower when compared to a direct call to the C code, but the interface between ikarus and libffi (all the steps performed from the point you call an ffi procedure from your Scheme code to the point where control reaches the C destination) would be the major contributor if I were to guess. > but can something be done about it? This needs some investigation. Ikarus's ffi was written such that I would understand all caveats involved in writing it; performance tuning would've made understanding harder. The implementation of the ffi may need a complete rewrite. Need to think more about it. Aziz,,,
