CCA Query Performance

Can someone please tell me why everytime we run a query using the 0Occa_c11 cube and I watch the query run using SM50 it always does a long sequential read of the cube fact table. Its seems like it retrieves all the data at the database level ignoring the fiscal year constraint and then filters the data on the client instead of on the database server for fiscal year. Can someone please help out. I need to make this query run faster.

Thanks

Richard