>Again, I'd suggest running your tests (Solaris; part vs non-part) with *ALL* options (showplan, stats >io, stats time, etc) turned off. The objective is to eliminate the overhead (cpu, logicalIOs) required to >process the enabled options.
We did tests with all options diabled, there is no real measurable change against the version with "set statistics time,io on". Attached is an output for the partitioned table. First with options "set statistics time,io" on, the second run with all options off.
>If performance is still degraded after disabling all those options, could you post back here with the >output from the following when run against the Solaris and linux dataservers:
Solaris no use of statement cache:
Parameter Name Default Memory Used Config Value Run Value Unit Type
------------------------------ -------------------- ----------- -------------------- -------------------- -------------------- --------------------
statement cache size 0 0 0 0 memory pages(2k) dynamic
enable literal autoparam 0 0 0 0 switch dynamic
Linux:
Parameter Name Default Memory Used Config Value Run Value Unit Type
------------------------------ -------------------- ----------- -------------------- -------------------- -------------------- --------------------
statement cache size 0 176474 80384 80384 memory pages(2k) dynamic
enable literal autoparam 0 0 1 1 switch dynamic