Pass max_dp3_threads to dp3_prep_targ.cwl in prep.cwl

In the HBA target workflow, the max_dp3_threads parameter was not passed to steps/dp3_prep_targ.cwl when it is called in worfklows/linc_target/prep.cwl and the former has no default value set. This caused it to launch as many prep jobs as my machine had cores, all trying to use all cores, even though I had max_dp3_threads=10 in the input parameters. resulting in >700% system load. This MR passes along the max_dp3_threads parameter to prevent this from happening. On my 96 core machine it now restricts itself to 9 simultaneous prep jobs.