Re: Query regarding use of benchmarks in documentation

On Sun, Jan 10, 2016 at 12:36 AM, Matt (Soul Studios)wrote:
> For libraries where real-world performance is a pragmatic consideration,
> does putting comparative benchmarks in the documentation make sense? I ask
> because I do not see any in my brief reading of the boost docs, but
> obviously complexity guarantees are not sufficient (or even relevant) in
> this way.

Boost library documentation includes benchmarks or other performance
characteristic information if the authors choose. For example:
http://www.boost.org/doc/libs/1_60_0/libs/context/doc/html/context/performance.html
http://www.boost.org/doc/libs/1_60_0/libs/smart_ptr/smarttests.htm
https://olk.github.io/libs/fiber/doc/html/fiber/performance.html

Even proposed libraries that want to be considered for inclusion in
Boost provide them. For example:
http://det.github.io/segmented_tree/boost_segmentedtree/performance.html

If you are submitting a library for Boost, and performance is
important, (and especially if you claim your implementation has good
performance) include benchmarks in your documentation.

Glen