[zeromq-dev] Parallel sort

Pieter Hintjens ph at imatix.com
Mon Aug 16 01:10:10 CEST 2010

On Mon, Aug 16, 2010 at 1:02 AM, Martin Sustrik <sustrik at moloch.sk> wrote:

> One option would be to rewrite the sort in iterative (rather than recursive)
> manner and apply divide and conquer approach on the result.

Any recursive algorithm can be turned into an iterative one.  You push
to a stack rather than executing, and then you execute one stack
dataset at a time.


More information about the zeromq-dev mailing list