There are some questions in the world that are mathrelated and still unsolved.
Which of these that are "hardest" to solve is virtually impossible to decide.
For a mathquestion to be hard, it must not be impossible.
In order to distinguish between hardest and impossible, one need to know an answer.
Often a mathproblem can be very hard to solve but even virtually impossible to prove that it is solved correctly.
One of the really hard mathproblems is Fermat's last theorem.
Fermat's Last Theorem states that
has no non-zero integer solutions for x, y and z when n > 2.
We were well into 1995 before there was a solution or a proof for this theorem.
Pierre Fermat died in 1665, so the theorem was without proof for 330 years.
It can however be disputed wether this is a math question.
I would think that to calculate the length of ALL roads combined in the world is a real challenge.
To calculate the individual movement of atoms in a microprosessor would be virtually impossible but even for ONE single transistor this is a huge task that with powerfull computers takes hours to complete.