Performance issue - CLRLIB vs RNMOBJ vs DLTLIB / Timeslice

We currently submit a job to copy data libraries from our prod environment to a 
test environment.  The libraries are very large, and we are currently doing a 
CLRLIB of the to-library followed by a CPYLIB from the prod library to populate 
the cleared test library with production data.  The job takes longer than we 
would like for it to complete, about 2 hours.

With large libraries might another approach perform better?  We thought that 
RNMOBJ/CRTLIB/CPYLIB might help the main process complete faster, allowing the 
DLTLIB of the unneeded RNMOBJ library to be handled later, but we didn't notice 
the job completing any faster.  I would think that a DLTLIB/CRTLIB/CPYLIB 
combination might be even slower in completing.

Have you been there done that?   

One of the first steps in the CLP is a CHGJOB to set the timeslice to 
6,000,000.  Someone once stated that too high of a timeslice will work against 
your own job performance-wise.  Any comments, suggestions?

Thanks,
Steve