|
Boost : |
From: Dietmar Kuehl (dietmar_kuehl_at_[hidden])
Date: 2002-05-23 18:39:00
Herve Bronnimann wrote:
> That's pretty much everybody else. And you can actually see the
> difference in terms of performance.
The difference in terms of performance is, however, probably not due
to the use of inferior allocators but rather to the more involved
maintenance of the objects: The need to identify the objects whose
key is to be modified basically means that they have to stay in place
in one form or the other. This in turn means that you have to modify
several pointer when rearranging the heap.
The node based heaps would probably benefit from a fast allocator, eg.
one for a fixed small size of objects drawn from a pre-allocated arary,
but I wouldn't expect a performance improvement of more than 10%. I
haven't measured the difference, however. Maybe I'm wrong and the major
time consumer is indeed the allocation...
-- <mailto:dietmar_kuehl_at_[hidden]> <http://www.dietmar-kuehl.de/> Phaidros eaSE - Easy Software Engineering: <http://www.phaidros.com/>
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk