Fix O(N^2) operation in Material

[MicahGale]

          My thoughts so far on the these results.

1. I realized a very bad implementation detail: In MCNP_Probelm.update_pointers there is a for data_input in self.data_inputs loop effective.
   1. This then calls Material.update_pointers which has: for data_input in data_inputs: loop.
   2. That's right: I wrote an O(N²) operation! 🎉 🎉 I found the worst implementation to do.
   3. This operation is linking M to MT. If we did vice versa MT could use hashing to move to an O(1) operation.
   4. This is an easy fix
2. Why is Material.__hash__ getting called so much? NumberedObjectCollection is meant to hash on numbers and not objects I thought. There is something weird going on here that needs further investigation.

[MicahGale]

See also: #137 (comment)

[MicahGale]

Some new profiling data:

1    0.000    0.000  272.390  272.390 /opt/hostedtoolcache/Python/3.12.5/x64/lib/python3.12/site-packages/montepy/input_parser/input_reader.py:6(read_input)
        1    0.026    0.026  272.390  272.390 /opt/hostedtoolcache/Python/3.12.5/x64/lib/python3.12/site-packages/montepy/mcnp_problem.py:236(parse_input)
        1    0.002    0.002  253.773  253.773 /opt/hostedtoolcache/Python/3.12.5/x64/lib/python3.12/site-packages/montepy/mcnp_problem.py:304(__update_internal_pointers)
     1001    0.363    0.000  252.7[21](https://github.com/idaholab/MontePy/actions/runs/10586963944/job/29336774706?pr=513#step:7:22)    0.252 /opt/hostedtoolcache/Python/3.12.5/x64/lib/python3.12/site-packages/montepy/data_inputs/material.py:173(update_pointers)

So this is very very inefficient. I think if this were fixed we could open the benchmark case in ~20 seconds?