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: 2013-10-24 01:59:03

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

Comment (by aiapub-cpp@…):

 I attached diagrams showing memory usage with and without the mix
 function. You can see that memory is used differently - the first one has
 longer bucket-chains. Peak memory usage is the same for both - 4GB.

 And you were right about the histograms (the time is in seconds):

 With mix code:

 {{{
 Fill: boost::unordered_set (64K)        - 00:52
 Bucket count: 134217728
 0 63716044
 1 47468076
 2 17685158
 3 4390091
 4 819351
 5 122072
 6 15183
 7 1564
 8 176
 9 13

 Fetch: boost::unordered_set (64K)       - 00:15
 Fetch: boost::unordered_set non-existing records(64K)   - 00:18
 Total time: 01:41
 }}}


 Without the mix code:

 {{{
 Fill: boost::unordered_set (64K)        - 00:07
 Bucket count: 134217728
 0 34217728
 1 100000000

 Fetch: boost::unordered_set (64K)       - 00:01
 Fetch: boost::unordered_set non-existing records (64K)   - 00:00
 Total time: 00:09
 }}}

-- 
Ticket URL: <https://svn.boost.org/trac/boost/ticket/9282#comment:6>
Boost C++ Libraries <http://www.boost.org/>
Boost provides free peer-reviewed portable C++ source libraries.