Adam Wulkiewicz via Boost-users <boost-users@lists.boost.org> writes:

> W dniu 14.06.2021 o 12:26, Jens Thiele pisze:
>> Adam Wulkiewicz via Boost-users <boost-users@lists.boost.org> writes:
>>
>>> Without knowing your code I can't esstimate if this is long time or
>>> not, I can only guess. In general qbegin() loop gathering some number
>>> of nearest elements should take similar time to query() call doing the
>>> same. So you can try getting some number of knn boxes from the R-tree
>>> and checking if this is the case.
>> it looks like I have some fundamental misunderstanding here:
>> t->qbegin(idx::nearest(pt, 100)) is really fast (<1ms) but
>> t->qbegin(idx::nearest(pt, tree_size)) with tree_size=37133 is really
>> slow (~500ms)
>>
>> shouldn't those calls take approximately the same time? I thought the
>> search is done incrementally?
>
> Yes, this doesn't look right. I have to look into this.
> Would you mind creating an issue at github?

https://github.com/boostorg/geometry/issues/867

Jens
_______________________________________________
Boost-users mailing list
Boost-users@lists.boost.org
https://lists.boost.org/mailman/listinfo.cgi/boost-users

Reply via email to