[Place] Unnecessary microoptimization in override_delay_model

[AmirhosseinPoolad]

Back in 2020 compilers weren't as good as they are now and t_override::operator< generated bad code when using std::tie (See #1225 ), which resulted in code that was faster but is not very easy to read and understand. This seems to no longer be the case anymore and gcc-13 successfully inlines t_override::operator< with std::tie.
https://godbolt.org/z/z8W8jxqqe

If you want to tackle this, you can simply copy paste the implementation in the above link and run vtr and titan benchmarks, compare them with master and make sure the generated assembly isn't actually slower.