Le vendredi 29 octobre 2010 à 10:51 +0200, M.-A. Lemburg a écrit :
I've thought about that. The main annoyance is to be able to switch transparently between the two implementations. But I think it would be interesting to pursue that effort, since indeed dicts with interned keys are the most common case of dicts in the average Python workload. Saving 1/3 of the memory size on these dicts would be worthwhile IMO.
Are you sure ? In the age of GB RAM, runtime performance appears to be more important than RAM usage. Moving the hash comparison out of the dict would likely cause (cache) locality to no longer trigger.
Good point. It probably depends on the collision rate. Also, a string key dict could be optimized for interned strings, in which case the hash comparison is unnecessary. (knowing whether the key is interned could be stored in e.g. the low-order bit of the key pointer) Regards Antoine.