In concolic testing, what does "concrete execution" mean?

In the context you've mentioned, I'm pretty sure that "concrete execution" refers to actually running the program on specific inputs and seeing what happens. The "concolic testing" article you've linked to suggests a hybrid approach between testing on specific inputs (concrete execution, which is complete but unsound) and symbolic testing (symbolic execution, which sound but incomplete).

Hope this helps!

Concolic execution is a mix between CONCrete execution and symbOLIC execution, with the purpose of feasibility.

Symbolic execution allows us to execute a program through all possible execution paths, thus achieving all possible path conditions (path condition = the set of logical constraints that takes us to a specific point in the execution). The problem is that, except for micro benchmarks, the cost of executing a program through all possible execution paths is exponentially large, thus prohibitive.

On the other hand, if we provide the symbolic execution with concrete values, you can guide it through a specific execution path (without traversing all of them) and achieve the respective path condition. This is feasible.

I hope this answers your question