Investigate graph evaluation for optimization opportunities

Evaluation of big graph solutions has been mentioned as a pain pint by several customers.
Although we have already done investigation of evaluation optimization, with no easy opportunities found, evaluation in graph build scenario is different in that it evaluates all projects to build graph.
Additionally, since graph build prohibits dynamic graph shaping (like for example conditional project dependency based on Target result) there might be good caching opportunity.

This is time boxed research producing artifacts:
- document current performance characteristics and compare to non graph
- prioritized list of optimization opportunities

Above information will be used for discussion with related SME.

List of related issues:
- https://github.com/dotnet/msbuild/pull/9083
- https://dev.azure.com/devdiv/DevDiv/_workitems/edit/1834963/