I have a general question about performance of Gurobi solver in light of cores vs. threads. Suppose that you have choice to solve MILP model on [8 cores / 8 threads] or [6 cores / 12 threads] processor. Could you give me some insight into which one would you choose? I suppose [8 cores / 8 threads] should be better for solver, but I'm not quite sure.
Please sign in to leave a comment.