Add benchmarks to prevent major performance regressions

[ashton314]

We'd like to make sure FloatTracker doesn't get too slow. We should have some simple tests/benchmarks to make sure we're not murdering our performance with a change.

[bennn]

Let's start with the examples we already have. Make a script that runs them all once & records the time taken.

EDIT: FPBench should be good for this. We can run the kernels on a bunch of sample of inputs. https://fpbench.org/benchmarks.html

EDIT: here are a bunch from FPBench https://github.com/utahplt/FloatTrackerExamples/tree/main/examples/FPBench