xufei
xufei
> what's the overall performance gain when running some SQL with hash job? you mean hash join v2 vs hash join v1 or hash join v2 with tagged pointer vs...
> > > what's the overall performance gain when running some SQL with hash job? > > > > > > you mean hash join v2 vs hash join v1...
> > which I think can effectively reduce the CPU time of hash join > > any compare result about this part? I personally prefer other ways of optimization to...
> > So I think if we can reduce the probe collision, it is for sure that we can reduce the cpu time. > > Thanks for the detailed explanation....
> > So I think if we can reduce the probe collision, it is for sure that we can reduce the cpu time. > > Thanks for the detailed explanation....
@winoros I can reproduce this bug. I think there is 2 unexpected thing: 1. the hash join use `CARTESIAN inner join`, but it actually has an equal join condition ```...
/test unit-test
/test pull-integration-ddl-test
/test check-dev2
/test pull-mysql-client-test