If you mean you are using torque or openPBS, then you should submit your job to a node and define the amount of cores you want to use.
qsub -l nodes=x:ppn=y jobname
this should do the trick.
x is the amount of nodes you want to use, y is the amount of cores you want to use on those nodes.
If you only intend to use OpenMP and no cluster openmp of intel or MPI or things of that sort, you should only do
qsub -l nodes=1:ppn=y jobname