On 06/02/15 01:47, Josh Berkus wrote:
> On 06/01/2015 03:22 PM, Tomas Vondra wrote:
>>
>> On 06/01/15 23:47, Josh Berkus wrote:
>>> On 06/01/2015 02:18 PM, Tom Lane wrote:
>>>
>>>> Anybody else want to speak for or against back-patching the patch as
>>>> posted? I intentionally didn't push it in before today's releases,
>>>> but I will push it later this week if there are not objections.
>>>
>>> I would like Mark Wong to test this on DBT3, if that's possible
>>> for him. I'm very worried about an unanticipated regression.
>>
>> AFAIK Mark is busy with other stuff at the moment, but I can do
>> the TPC-H (which DBT3 is equal to, IIRC).
>
> Yeah, I just want something which has a chance of catching
> unanticipated regression in queries which should be unaffected by the
> patch.
OK, so I did the testing today - with TPC-H and TPC-DS benchmarks. The
results are good, IMHO.
With TPC-H, I've used 1GB and 4GB datasets, and I've seen no plan
changes at all. I don't plan to run the tests on larger data sets, I do
expect the behavior to remain the same, considering the uniformity of
TPC-H data sets.
With TPC-DS (using the 63 queries supported by PostgreSQL), I've seen
two cases of plan changes - see the plans attached. In both cases
however the plan change results in much better performance. While on
master the queries took 23 and 18 seconds, with the two patches it's
only 7 and 3. This is just the 1GB dataset. I'll repeat the test with
the 4GB dataset and post an update if there are any changes.
While this can't prove there are no regressions, in these two benchmarks
the patches actually improve some of the queries.
regards
--
Tomas Vondra http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
