feat(pumpkin-core): Add additional (coarse) timing statistics

[EmirDe]

In preparation for larger-scale profiling.

Adding measurements so that we can see how much time is being spent in:

• Fixpoint propagation
• Conflict analysis, and the two nogood minimisation mechanisms
• Branching

This should be inexpensive to measure since we only record this once per fixpoint propagation call or conflict analysis.

I will later open a separate PR where we do a bit more detailed statistics per propagator.

Now this is a draft PR to see if there are any issue, overall it seems ready to me, but I have yet to run it on some instances.

[ImkoMarijnissen]

Overall, the current changes look good to me.

I recall that when I was doing fine-grained profiling when we were rewriting the solver that the time spent in measuring could be quite influential on performance, especially when there were many decisions/failures, so it would be good to profile!

We could consider using something like minstant or coarsetime if it turns out to be a bottleneck.

[EmirDe]

Overall, the current changes look good to me.

I recall that when I was doing fine-grained profiling when we were rewriting the solver that the time spent in measuring could be quite influential on performance, especially when there were many decisions/failures, so it would be good to profile!

We could consider using something like minstant or coarsetime if it turns out to be a bottleneck.

For fine-grained profiling, I made it so that it is not run by default, and instead it is a compile-time flag that needs to be set, similar to how we do it for checking propagations. I want to run some tests to see how often to particular propagators run and how often they propagate (but this is for another PR).