RE: Is hash join slower in 12.2 compared to 11.2.0.3 or am i hitting a bug?

Hello Nenad,

> kdstf111010100001000km is a dynamically generated function doing full table scan.

Yes, that is what I tried to say. In fact it is a little bit more sophisticated. kdstf*km and the following bitmap numbers are an implementation to reduce branch mispredictions (CPU execution) and so on.

Best Regards
Stefan Koehler

Independent Oracle performance consultant and researcher Website: http://www.soocs.de
Twitter: _at_OracleSK

> Noveljic Nenad <nenad.noveljic_at_vontobel.com> hat am 27. Januar 2020 um 09:27 geschrieben:
>
> Hi Stefan,
> > As already mentioned the stacks are driven by top function clock_gettime() - however the calling functions are not particularly related to the HJ/FTS (e.g. kdstf111010100001000km().
>
> kdstf111010100001000km is a dynamically generated function doing full table scan.
> > What are these functions (e.g. 000000002D93D64B, 0000000014D30516, etc.) without any symbols about?
>
> This is probably dynamically ("just-in-time") generated code presumably sitting in the private memory. A longer stack would provide more context.
>
> Best regards,
>
> Nenad
>
> https://nenadnoveljic.com/blog/

--
http://www.freelists.org/webpage/oracle-l