When was that? Before or after the last protocol fork (which made it efficient with most ASICs) which happened in July? If it was after the fork it's surprising as it should have solved these problems.
It was running 13.3. I'll maybe have another go with it sometime, perhaps there was something strange with my install, or I missed something. I'm pretty sure I did everything I should have, though.
I'm rejiggering my workstation in the next few days, including a fresh install, so I'll see if I can have another go.
Still, all that said, I'm still getting far better returns from a larger pool than I did on p2pool.
Kinda a catch 22 there... More hash rate == Quicker Block solves == lower variance. (PPLNS payout) p2pool has less total hash rate so therefore it will have higher variance which will lead to fewer people using it which leads to lower total hash rate....