Investigate Fluxion match performance with different policies, graphs, subsystems

[milroy]

Issue #1001 identified helped identify performance problems with Fluxion matching and scoring on Hetchy. We need to perform a comprehensive study of Fluxion performance under different:

• simulated current (O(1000) nodes) and next-generation (>O(10,000) nodes) clusters with representative heterogeneity
• subsystems
• realistic jobspecs/workloads
• queue policies
• match policies
• scoring

The study will help us understand and prioritize implementing performance improvements.

[tpatki]

Maybe we can use PerfFlowAspect for this. Adding @slabasan as well, we had an initial integration of fluxion and PFA that maybe we can start with