[Python-Dev] Evil reference cycles caused Exception.traceback (original) (raw)

francismb francismb at email.de
Mon Sep 25 13:51:29 EDT 2017

Hi, just curious on this,

On 09/18/2017 10:54 PM, Antoine Pitrou wrote:

I'm not an expert on GC at all, but intuitively it sure seems like allocation size might be a useful piece of information to feed into a heuristic. Our current heuristic is just, run a small collection after every 700 allocations, run a larger collection after 10 smaller collections. Ok, so there are two useful things a heuristic may try to guess: 1. how much CPU time a (full) collection will cost 2. how much memory it might help release While #1 is quite realistically predictable (because a full collection is basically O(number of allocated objects)), #2 is entirely unpredictable and you can really only make out an upper bound (trivially, the GC cannot release any more memory than has been allocated :-)).

is the allocation size not useful enough for #2 ? (the upper bound seems logical as far as the location is done by the same "entity", (or size communicated back to) )

Thanks in advance! --francis

More information about the Python-Dev mailing list