Db_file_multiblock_read_count

[slimdave]

I respectfully disagree.

This parameter greatly dependes upon the type of system you are setting up. If OLTP/DSS then a smaller number in the 4-16 is probably a very good choice.

But if its say a DW, then large number up to the maximim I/O size would be a better choice.

Also, your block size should be greater than or equal to the OS I/O size.

Regards,
Ken

If you're reducing the MBRC to tilt the optimizer away from table scans then you ought instead to consider either gathering system statistics (to benchmark the relative performance of single and multiblock reads) or set optimizer_index_cost_adj to give roughly the same info.