I tried my
P4 direct die block on a normal p4 (IHS still attached) with good results, unfortunatly I couldn't compare temps since both the cpu and mobo we're different (and we know how accurate cpu temps are anyway...)
I don't know about actually drilling cups into the IHS, but divots shouldn't be a problem to do , personally I'd simply attack the IHS with a brass wire wheel to roughen the surface up, that way you simply have to lap it to enable re-use of "traditional" cooling later on.