[Benchmarks] Refactor CPU count scenarios results parsing

### Is your feature request related to a problem? Please describe

With PR https://github.com/intel/llvm/pull/19857, for some `sycl` scenarios of Compute Benchmarks we know from the beginning that the scenario's result type is either time or cpu instructions count.

Currently, the results are grouped into CPU count charts based on parsing the results file and extending the scenario's naming in metadata only after benchmark finishes.

Refactor the code so that all scenarios measuring CPU count are labeled differently than the same scenarios that measure time. This would simplify the code, removing the need for determining the results' chart group based on a log parse. This can require introducing a refactor in Compute Benchmarks repo for `ur` benchmarks to use a new functionality of the `Combo Profiler`, see `sycl` scenarios changes: https://github.com/intel/compute-benchmarks/commit/89119f2f5ccb24be36fa616fe5d945cb14a8b2c9.

### Describe the solution you would like

_No response_

### Describe alternatives you have considered

_No response_

### Additional context

_No response_