Aggregation is not being designed for large dimension attribute, although aggregation usage property is set to Unrestricted

I have a large dimension having more than 10 million records. The dimension has two attributes. I have set the AggregationUsage property to "Unrestricted" for the non key attribute. But when I query the cube against this particular dimension, I find that partitions are being scanned instead of aggregations being used. I believe due to such huge number of records, aggregation wizard is not designing aggregations for this particular attribute. Can I somehow force the wizard to design aggregations for this attribute ?

Note -  I'm using SSAS 2008.