It is only procedures with 'SP_' prefix that cause the performance impact of scanning the system databases first.
'SP' should be fiine.
So, isn't some similar scanning occur when all the stored procedures start with the same prefix? ('SP' in this case). Not maybe for the system databases, but within the procedures in that specific context?
Although I have not tested it properly, it seems that one should name his/her stored procedures as unique as possible for the maximum performance, starting from the first character.
Were any tests done on that?