if this function is taking a unusual amount of time compared to your expectations, the system we're using is CW's standard HGX node, with a Xeon Platinum 8462Y+ , 2 TB ram