There are a few things I noticed when writing this that I think we should fix:

* `passed`: Boolean indicating if verification passed (default: `True`).

We should just throw exceptions when a benchmark fails, not produce an invalid result.

* `name`: Set to `label` by the framework.

The name field in Result is inconsistently used throughout the framework. We should clean this up and always use label where appropriate.

* **Use unique names:** Ensure `benchmark.name()` and `result.label` are descriptive and unique.

What I was contemplating writing is that benchmark names should ideally never change because they are unique keys the visualization scripts use to collect benchmark data into series.
But this is a very annoying limitation. We should introduce a (maybe autogenerated?) key to groups/benchmarks that is unique and not used for any sort of visualization (or maybe treat the existing name/label as a key, and add a new "display name"?)

What I was contemplating writing is that benchmark names should ideally never change because they are unique keys the visualization scripts use to collect benchmark data into series. But this is a very annoying limitation. We should introduce a (maybe autogenerated?) key to groups/benchmarks that is unique and not used for any sort of visualization (or maybe treat the existing name/label as a key, and add a new "display name"?)

I think that new "display_name" and "display_label" variables would work best in this situation. They could by default take the current value of name/label and be overridden, if needed.

@intel/llvm-gatekeepers please merge