Fixes bug with nested OpenMP, fixed task threshold, extended tests range

[sterrettm2]

Fixes the bug with nested OpenMP by adding #pragma omp taskwait Changes the task_threshold when OpenMP is enabled but parallelization isn't chosen from 0 to the max value for arrsize_t; I don't think this was causing any problems, but it was definitely a mistake Additionally, extended the kvsort tests to also run for sizes of 10^4, 10^5, 10^6 for testing the OpenMP logic