[CI][Benchmarks] compute benchmarks charts naming

[PatKamin]

Improve readability of Historical Results chart titles, tooltips, and all charts' legends by introducing new Result object field: display_name for Compute Benchmarks

[pbalcer]

Here's an idea that might simplify this patch:
move display_name() to be a benchmark metadata field. That we we never have to update old results, and we can just look it up in metadata in the scripts if needed.

[PatKamin]

Here's an idea that might simplify this patch: move display_name() to be a benchmark metadata field. That we we never have to update old results, and we can just look it up in metadata in the scripts if needed.

Thanks, made changes

[pbalcer]

mostly lgtm, I like this approach better. I'm not sure what we can do with CPU Count stuff, but I really don't like fixing the metadata after the benchmarks ran.
If we can't figure out a better approach, we should try to isolate this hack better.

[pbalcer]

can you show a screenshot with how the new display names look like (on all types of charts)?

[PatKamin]

can you show a screenshot with how the new display names look like (on all types of charts)?

Here are few examples, one for each type of charts (grouped results have updated legends, Historical Results charts - titles, for grouped results titles there will be another PR):

[pbalcer]

awesome, thanks!

[pbalcer]

@intel/llvm-gatekeepers please merge. The CI failure is unrelated.