Benchmarking interface for SOTA solvers

[phstratmann]

User story

As a user, I want to be able to easily run and benchmark QUBO problems on state-of-the-art solvers running on CPU. My goal is to quickly validate that Loihi 2 outperforms alternative solvers and to quantify this outperformance. This feature allows me to check if Loihi 2 is benefitial for my work load, and to create reproducible numbers for sharing with other stakeholders of a project.

Conditions of satisfaction

• User can run their Lava-defined QUBO workload in QBSolv and Gurobi.
• Lava returns time to solution for SOTA solver.
• Lava allows to activate/deactivate running the solver, to measure energy consumption with and without solver routine.

Acceptance tests

• Test defining an MIS as QUBO class and running it in QBSolv.
• Test defining an MIS as QUBO class and running it in Gurobi.
• Test running an MIS in QBSolv and returning the time to solution.
• Test running an MIS in Gurobi and returning the time to solution.
• Test running an MIS in QBSolv and deactivating the actual call function.
• Test running an MIS in Gurobi and deactivating the actual call function.