If you want the Rosetta@Home CPID (the older CPID) to transfer to the WCG CPID, you have to run both projects on BOINC. This way, they know that the older CPID should apply to the WCG. Usually, it only takes a few hours for the update to occur. This is from the BOINC GUI, I'm not sure if the boinccmd (command line version on Linux) knows if other project's CPID should be changed.
Thx, that's what I'm currently doing right now, let's see how things work out.
Additionally I copied my desired CPID to the "external_cpid" field in client_state.xml (as described here: https://steemit.com/gridcoin/@dutch/the-definitive-guide-to-fixing-a-split-cpid-gridcoin-s-and-boinc-s-1-problem). Not sure if this is gonna help^^.
On a completely unrelated note: I think someone found a way to crunch Rosetta tasks with GPUs. This user: https://boinc.bakerlab.org/rosetta/hosts_user.php?sort=expavg_credit&rev=0&show_all=0&userid=1994560 has machines with completely ridiculous CPUs but sporting 6 AMD cards. His best machine is currently at a whopping 19k RAC, which most certainly does not hail from the single Celeron G1620.
I read that split cpids article but I think they overcomplicated the issue. (Ive had a lot of servers in the past running various boinc projects and sold a lot of these) and it all really boils down to correcting the e-mail address and worst case scenario: Running RAH first, unmounting WCG and re-mounting WCG. Ive never seen the need to actually edit the XML file, wcg does the heavy lifting in the background.
I just checked the email hash; it's the same for both projects...
Edit: Ah well, so now I tried REMOVING RAH from the machine where I attached WCG, then updating WCG. As a result this has now correctly inserted my desired CPID into the field "external_cpid" in the XML file. Now I will wait until this also shows up on the WCG web page/profile; on the other machine the "external_cpid" is still empty...