What are you parms set too :
Degree of parallelism (DFT_DEGREE) =
Maximum query degree of parallelism (MAX_QUERYDEGREE) =
Enable intra-partition parallelism (INTRA_PARALLEL) =
You need to make sure you have atleast 1 cpu for each MAX_QUERYDEGREE,
One of the things we noticed on loads is, while writing it would wait when a bunch of procesess were doing loads at the same time, we now set it at
DFT_DEGREE = 1
MAX_QUERYDEGREE = 4
INTRA_PARALLEL = YES,
Our loads now go quicker and it will use more procesess to build indexes.