Why does only one cpu get used with my SQL 6.5 system for an extremely complex/long running query?

A. This is because under SQL 6.5 only a single cpu can be used for a query. The only time another cpu is used is if read-ahead is being done and this can then use another cpu.

Under 7.0 all cpu's can potentially be used for a single query if that query can be parallelised.