When using scipy RBF interpolation routines it seems that all the cores are used by default.
This is ideal to speed up the computation but not ideal on machines where there are other jobs running and the resources are shared.
Is there any way to limit the number of CPU cores available to scipy for computation when interpolating?