More reasonable and accurate benchmarks

[lemire]

Currently, the CRoaring project does not have, by itself, sensible microbenchmarks. This PR will fix that.

(The 'benchmarks' directory is filled with ad hoc benchmarks that may have their uses.)

Running microbenchmarks

We have microbenchmarks constructed with the Google Benchmarks.
Under Linux or macOS, you may run them as follows:

cmake --build build
./build/microbenchmarks/bench

By default, the benchmark tools picks one data set (e.g., CRoaring/benchmarks/realdata/census1881).
We have several data sets and you may pick others:

./build/microbenchmarks/bench benchmarks/realdata/wikileaks-noquotes 

You may disable some functionality for the purpose of benchmarking. For example, you could
benchmark the code without AVX-512 even if both your processor and compiler supports it:

cmake -B buildnoavx512  -D ROARING_DISABLE_AVX512=ON
cmake --build buildnoavx512
./buildnoavx512/microbenchmarks/bench

[huihan365]

Glad to see we have more accurate benchmark. I will use it.

[lemire]

@huihan365 We are often bound by array containers. Use perf record/perf report to identify the bottlenecks: you can require the benchmark to focus on on operation with --benchmark_filter=.....

Now that we have the foundation for AVX-512, we should be able to really optimize it well.

[huihan365]

Got it, i will work on it when i finish my work on-hand.