If the problem was the number of datamapping engines, you wouldn’t see the one used being slower…jobs would simply be queued.
From what you have explained, it seems like the resources are greatly used while a big job goes through which seems to leave less for other processes, like the datamapper.
When you compare both scenario (with and without Content Creation), are you using the same data file (that is in use by the Datamapper) and same output job (being in use by the Content Creation)?
Also, do you have a anti-virus which is scanning the jobs? If so, can you turn it off for testing purposes?
While it is going slow (as we have seen, it can take up to 40 minutes), can you look at the Windows Task Manager and see where the resources are being used?
As for your question about the number of Datamapping engines, if you have a lot of processes calling the Datamapper engines simultaneously, having more engines makes sense. You might want to increase the memory of the Datamapping engines has they could require more “juice” when dealing with big data file.
As for the memory allocated, it all depends of the number of simultaneous jobs and their size. You might try to increase it but remember that your OS and other software running on your server do need them as well.
I whish their was an equation to calculate the best performance for each user but their is none. It varies depending of so many factors that it is trial and error to find the proper one for you. Then again, should your job profile change a lot (number of small, medium and big jobs) running simultaneously, then your settings will need to change as well.
Also, please note that with Connect 2020.1, the Connect Server Configuration has been improved when it comes to the engines configuration. It has been made more simple and the default value represent more and more the “usual” environment use of customer…then again, not all are the same.
Hope that guide you a little more.