SAT solver optimizations of DPLL for unsatisfiable formulas [duplicate]

Question

This question already has an answer here:

• Best improvements to do to the DPLL SAT algorithm 2 answers

I am implementing a SAT solver based on DPLL algorithm, and it works fine on small formulas and larger satisfiable problems. My case split is based on a tree like structure, where every branch is tried one by one until solution for the problem is found or until every branch is tried.

So my problem now is that the above described approach does not terminate when trying to find a solution on large UN-SAT problems (so there shouldn't be a solution), therefore my program tries every possibility $2^n$, giving it a run time of years to solve UN-SAT problem with $50$+ variables (it would eventually terminate in years as I am not having any memory issues etc.).

My question is: what topics/implementations would I want to look at to speed up the tree processing or this case split process on unsatisfiable problems. Thanks.