[boost] [spinlock] concurrent_unordered_map is ready for review

Subject: [boost] [spinlock] concurrent_unordered_map is ready for review
From: Niall Douglas (s_sourceforge_at_[hidden])
Date: 2014-09-01 09:38:55

Dear Boost,

CC: Howard, Jonathan for N3645 enhancements

It has taken me far longer than I expected, however I believe that 
proposed boost::concurrent_unordered_map is ready for community 
feedback, with its unit test suite passing all green on the thread 
sanitiser and 92% coverage including exception safety tests. The main 
documentation page with the API and a lengthy explanation of the 
caveats in this design can be viewed at 
https://ci.nedprod.com/view/Boost%20Thread-Expected-Permit/job/Boost.S
pinlock%20Test%20Linux%20GCC%204.8/doxygen/classboost_1_1spinlock_1_1c
oncurrent__unordered__map.html

Performance improves linearly to CPU cores at a slope of about 0.9. I 
only have four core hardware available to me, if someone could tell 
me how it scales to eight cores or more I would be very interested.

Source code is at https://github.com/ned14/boost.spinlock. Tested on 
GCC 4.8, clang 3.4 and VS2013.

Niall

-- 
ned Productions Limited Consulting
http://www.nedproductions.biz/ 
http://ie.linkedin.com/in/nialldouglas/