Rebuild the stack of benchmarks

[jahofmann]

As part of the work in issue #79, the benchmarks that have been deprecated for a while have been removed as they were not functional anymore.

To evaluate the performance of TaPaSCo properly, again, there should be a replacement that checks the different performance critical parts of TaPaSCo.

There were four different tests:

• benchmark-alloc-free: Check memory allocation performance
• benchmark-cd: ???
• benchmark-latency: Counter execution latency
• benchmark-mem: Bandwidth for data movements -> Replaced by bandwidth tool

Additional tests should be added for the different parts of the architecture.

[cahz]

@jahofmann is this still relevant?

[jahofmann]

Sure, would still be nice to have more such tools to better catch regressions.