My take on the lit was that GRE (in general) was weakly to moderately correlated with graduate GPA. Though, this suffers from a restriction of range issues as many grad programs have a "less than B is a fail" policy, so there's not a lot of spread. I've seen conflicting data on prediction of completion of degree. Again, this is tough as, outside of diploma mills, attrition is a relatively rare thing. And, within that small N group, studies lump in those who were dismissed for academic reasons, and those who left for non-academic reasons. There was a Psych Bull study, a little dated by now I'm sure, that showed a relationship with GRE scores and research productivity. As for EPPP, I believe that was a Sharpless paper like a decade ago.