Thread: Patch Review Complete: Multi-Batch Hash Join Improvements

Note: this email is effectively a repeat of an email sent earlier to
which there has been less response than I expected. If there's something
else I'm supposed to do at this point, someone please let me know,
because I don't know what it is :)

-----------------------------------

I've finished all I think I can do to review the multi-batch hash join
improvement patch (http://is.gd/7W8Z). It appears to work as advertised:

-- it speeds up joins by fairly significant margins in some cases
-- I've been unable to make a case where it performs worse than the
existing implementation despite deliberately pathological inputs
-- it applies cleanly
-- it doesn't break stuff, AFAICS

I don't consider myself qualified to do a proper code review. I can only
suggest the author modify it to use /* ... */ style comments rather than
//... comments.

If there's more I should do, someone please let me know. Otherwise, I
leave it in the capable hands of the InterWebs.

- Josh / eggyknap

On Mon, 2008-11-17 at 23:19 -0700, Joshua Tolley wrote:
> -- it speeds up joins by fairly significant margins in some cases

The original claim in the message you cite says 10-50% for some data
distributions. Were you able to observe that kind of speedup?

Regards,Jeff Davis

On Mon, Nov 17, 2008 at 10:42:21PM -0800, Jeff Davis wrote:
> On Mon, 2008-11-17 at 23:19 -0700, Joshua Tolley wrote:
> > -- it speeds up joins by fairly significant margins in some cases
>
> The original claim in the message you cite says 10-50% for some data
> distributions. Were you able to observe that kind of speedup?
>
> Regards,
>     Jeff Davis
>

Oh, great, and now I can't find my numbers anymore :) After some
reflection I think there's some more formal verification I should expect
of myself before saying I'm done. I can say I saw results that supported
those claims, but I didn't gather enough data to prove them. Stay tuned
for details.

- Josh