You should also take in account your IO subsystem throughput which is a bottleneck on full table/index scans, parallel DML, etc. CPU is a limiting factor during sort, hash join, etc. For example, if you query single table with no joins than you might want to consider IO throughput as a limiting factor without taking in account number of CPUs (well, to a certain extent). As others mentioned, concurent parallel statements should be avoided (or accounted for). 2006/1/3, José Rodrigo <rodrigo@xxxxxxxxxx>: > Hi list, > > > When oracle start parallel query process by read init parameter > parallel_min_server he assing > a parallel query process to each processes in server? > My question is: > > An server with 1 processor can make user parallel query? > > Thank´s a lot. > > > -- > //www.freelists.org/webpage/oracle-l > > > -- Best regards, Alex Gorbachev -- //www.freelists.org/webpage/oracle-l