On benchmarks, heaps, priority queues

aaronwmail-usenet at yahoo.com aaronwmail-usenet at yahoo.com
Thu Jan 27 17:34:33 CET 2005



(re: http://xsdb.sourceforge.net/bench/pq3.py)

nsz> ...bisect is not so fast for large data...

Yes I know in theory the insertion sort approach should be bad for
large enough values, but the weird thing is that if you mix inserts and
deletes (with enough deletes) even 1M elements is not a large enough
value.  Anyway, for 10K or less the insertion sort priority queue
implementation seems to always be better.

Weird.  -- Aaron Watters

===
Hypothetical performance improvements are the root of all evil.
-- Bill Tutt (paraphrased)




More information about the Python-list mailing list