Remove warmup from benchmark

[Wsm2110]

Warmup will result in caching of delegates and compiling of all expressions. This will give containers with few features an unfair advantage over others.

I.e. warmup of 5 benchmarks will lead to for instance 20 cached entries. A warmup of every benchmark will lead to 70, resulting in more hashcollissions
This will cause benchmarks to interfere with others. Removing the warmup will cause benchmarks to be more accurate

[danielpalme]

I don't think it makes any sense to put further effort into this benchmark. It's outdated and has to be rewritten for .NET Core.

[Wsm2110]

Well, i guess you should archive this repo. it's outdated and the benchmarks are not accurate