> On 13 Sep 2024, at 15:01, Andreas Karlsson <andr...@proxel.se> wrote: > > On 9/10/24 10:54 AM, Daniel Gustafsson wrote: >>> On 22 Jul 2024, at 23:25, Andreas Karlsson <andr...@proxel.se> wrote: >>> >>> I have bench marked the two patches now and failed to measure any speedup >>> or slowdown from the first patch (removing return) but I think it is a good >>> idea anyway. >>> >>> For the second patch (optimize strict) I managed to measure a ~1% speed up >>> for the following query "SELECT sum(x + y + 1) FROM t;" over one million >>> rows. >> That's expected, this is mostly about refactoring the code to simplifying the >> JITed code (and making tiny strides towards JIT expression caching). > > Yup! Expected and nice tiny speedup. > >>> I would say both patches are ready for committer modulo my proposed style >>> fixes. >> I am a bit wary about removing the out_error label and goto since it may open >> up for reports from static analyzers about control reaching the end of a >> non-void function without a return. The other change has been incorporated. >> The attached v3 is a rebase to handle executor changes done since v2, with >> the >> above mentioned fix as well. If there are no objections I think we should >> apply this version. > > Sounds good to me and in my opinion this is ready to be committed.
This fell off the ever-growing TODO again. Re-reading it I still think it's a good idea, it applied almost cleanly still and still gives a slight performance improvement along with the more interesting refactoring which will make caching of expressions easier down the line. Attached is a rebased v4, unless there are objections I'll go ahead with this version. -- Daniel Gustafsson
v4-0002-Add-special-case-fast-paths-for-strict-functions.patch
Description: Binary data
v4-0001-Replace-EEOP_DONE-with-special-steps-for-return-n.patch
Description: Binary data
