Re: [Boost-bugs] [Boost C++ Libraries] #9282: performance regression in boost::unordered on 64-bit platforms

Subject: Re: [Boost-bugs] [Boost C++ Libraries] #9282: performance regression in boost::unordered on 64-bit platforms
From: Boost C++ Libraries (noreply_at_[hidden])
Date: 2014-02-24 18:23:10

#9282: performance regression in boost::unordered on 64-bit platforms
--------------------------+------------------------------------------------
  Reporter:  aiapub-      |      Owner:  danieljames
  cpp@…                   |     Status:  assigned
      Type:  Bugs         |  Component:  unordered
 Milestone:  Boost        |   Severity:  Regression
  1.56.0                  |   Keywords:  unordered mix64_policy performance
   Version:  Boost        |
  1.53.0                  |
Resolution:               |
--------------------------+------------------------------------------------

Comment (by anonymous):

 > This is slower in general, but this case is pretty important

 Agreed; prime tables (with MLF 1.0) will always outperform power-two
 tables with integral keys in the general case, since they produce no
 collisions. The only exception is if the hash function is guaranteed to be
 unique in the lower log_2(table_size) bits; in that case unmixed power-two
 hashing will win. Not sure where I would look to find such use cases
 (other than the toy insert-consecutive benchmark above), so I am
 enthusiastic for the above-referenced patch to go in.

-- 
Ticket URL: <https://svn.boost.org/trac/boost/ticket/9282#comment:10>
Boost C++ Libraries <http://www.boost.org/>
Boost provides free peer-reviewed portable C++ source libraries.