Tim Adye

OK, I guess I can't argue with that. Do these direct uses of `tbb` conflict with `tbbWrap`'s other function of bypassing `tbb` setup when running in a single-threaded mode? That...

> Empirical boundaries seem fine to me personally. I don't know if this is what @paulgessinger meant. Personally, I would prefer consistent boundaries, as I suggested in https://github.com/acts-project/acts/pull/3756#discussion_r1817850323 so you...

> I feat `constexpr` is only available with C++26 onwards for math functions. ah yes of course. I forgot I already mentioned that 😉

yeah, and the pix/str hole cuts aren't eta-dependent. Unfortunately the `Acts::TrackSelector` is a core class, so we can't add detector-dependent selections to it. I think that could be improved, but...

The ITk ttbar_pu200 efficiency etc are not significantly changed by this PR. I can't speak to the resolutions. However I do see a 4.7% speedup :) It's a long time...