topN using a heap
Andrei Alexandrescu via Digitalmars-d
digitalmars-d at puremagic.com
Sat Jan 16 19:26:54 PST 2016
On 1/16/16 9:37 PM, Timon Gehr wrote:
> Ivan's analysis suggests that even something significantly larger, like
> n/log(n)² might work as an upper bound for k.
I'm not clear on how you got to that boundary. There are a few
implementation details to be minded as well (quickly and carefully
approximating the log etc). Could you please make a commented pull
request against my own? Thanks! -- Andrei
More information about the Digitalmars-d
mailing list