Re: Gather Merge - Mailing list pgsql-hackers

On Mon, Nov 14, 2016 at 3:51 PM, Thomas Munro <thomas.munro@enterprisedb.com> wrote:

On Sat, Nov 12, 2016 at 1:56 AM, Rushabh Lathia
<rushabh.lathia@gmail.com> wrote:
> On Fri, Nov 4, 2016 at 8:30 AM, Thomas Munro <thomas.munro@enterprisedb.com>
> wrote:
>> + * Portions Copyright (c) 1996-2015, PostgreSQL Global Development Group
>> + * Portions Copyright (c) 1994, Regents of the University of California
>>
>> Shouldn't this say just "(c) 2016, PostgreSQL Global Development
>> Group"?
>
> Fixed.

The year also needs updating to 2016 in nodeGatherMerge.h.

Oops sorry, fixed now.
 

>> + /* Per-tuple heap maintenance cost */
>> + run_cost += path->path.rows * comparison_cost * 2.0 * logN;
>>
>> Why multiply by two?  The comment above this code says "about log2(N)
>> comparisons to delete the top heap entry and another log2(N)
>> comparisons to insert its successor".  In fact gather_merge_getnext
>> calls binaryheap_replace_first, which replaces the top element without
>> any comparisons at all and then performs a sift-down in log2(N)
>> comparisons to find its new position.  There is no per-tuple "delete"
>> involved.  We "replace" the top element with the value it already had,
>> just to trigger the sift-down, because we know that our comparator
>> function might have a new opinion of the sort order of this element.
>> Very clever!  The comment and the 2.0 factor in cost_gather_merge seem
>> to be wrong though -- or am I misreading the code?
>>
> See cost_merge_append.

That just got tweaked in commit 34ca0905.

Fixed.
 

> Looking at the plan I realize that this is happening because wrong costing
> for Gather Merge. Here in the plan we can see the row estimated by
> Gather Merge is wrong. This is because earlier patch GM was considering
> rows = subpath->rows, which is not true as subpath is partial path. So
> we need to multiple it with number of worker. Attached patch also fixed
> this issues. I also run the TPC-H benchmark with the patch and results
> are same as earlier.

In create_grouping_paths:
+                   double      total_groups = gmpath->rows *
gmpath->parallel_workers;

This hides a variable of the same name in the enclosing scope.  Maybe confusing?

In some other places like create_ordered_paths:
+       double      total_groups = path->rows * path->parallel_workers;

Though it probably made sense to use this variable name in
create_grouping_paths, wouldn't total_rows be better here?

Initially I just copied from the other places. I agree with you that

create_ordered_paths - total_rows make more sense.
 

It feels weird to be working back to a total row count estimate from
the partial one by simply multiplying by path->parallel_workers.
Gather Merge will underestimate the total rows when parallel_workers <
4, if using partial row estimates ultimately from  cost_seqscan which
assume some leader contribution.  I don't have a better idea though.
Reversing cost_seqscan's logic certainly doesn't seem right.  I don't
know how to make them agree on the leader's contribution AND give
principled answers, since there seems to be some kind of cyclic
dependency in the costing logic (cost_seqscan really needs to be given
a leader contribution estimate from its superpath which knows whether
it will allow the leader to pull tuples greedily/fairly or not, but
that superpath hasn't been created yet; cost_gather_merge needs the
row count from its subpath).  Or maybe I'm just confused.

Yes, I agree with you. But we can't really do changes into cost_seqscan.

Another option I can think of is just calculate the rows for gather merge,

by using the reverse formula which been used into cost_seqscan. So

we can completely remote the rows argument from the create_gather_merge_path()

and then inside create_gather_merge_path() - calculate the total_rows

using same formula which been used into cost_seqscan. This is working

fine - but not quite sure about the approach. So I attached that part of changes

as separate patch. Any suggestions?

--

Rushabh Lathia

www.EnterpriseDB.com