Effect of changing the value for PARALLEL_TUPLE_QUEUE_SIZE

From: Rafia Sabih <rafia(dot)sabih(at)enterprisedb(dot)com>
To: PostgreSQL Developers <pgsql-hackers(at)postgresql(dot)org>
Subject: Effect of changing the value for PARALLEL_TUPLE_QUEUE_SIZE
Date: 2017-05-30 09:28:04
Message-ID: CAOGQiiOAhNPB7Ow8E4r3dAcLB8LEy_t_oznGeB8B2yQbsj7BFA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox
Lists: pgsql-hackers

Hello everybody,

Here is a thing I observed in my recent experimentation, on changing
the value of PARALLEL_TUPLE_QUEUE_SIZE to 6553600, the performance of
a TPC-H query is improved by more than 50%.

Specifically, with this change, q12 completes in 14 seconds which was
taking 45 seconds on head. There wasn't any change in the plan
structure, just the time at gather-merge reduced which gave this

This clearly says that the current value of PARALLEL_TUPLE_QUEUE_SIZE
is not the best one for all the queries, rather some modification in
it is very likely to improve performance significantly. One way to do
is to give this parameters as another GUC just like
min_parallel_table_scan_size, etc.

Attached .txt file gives the plan at head and with this patch,
additionally patch is attached for setting PARALLEL_TUPLE_QUEUE_SIZE
to 6553600 too.

Rafia Sabih
EnterpriseDB: http://www.enterprisedb.com/

Attachment Content-Type Size
q12_ptqs.txt text/plain 3.5 KB
change_parallel_tuple_q_size.patch application/octet-stream 512 bytes


Browse pgsql-hackers by date

  From Date Subject
Next Message Fabien COELHO 2017-05-30 09:47:26 Re: pgbench more operators & functions
Previous Message Magnus Hagander 2017-05-30 09:26:00 Re: [COMMITTERS] Re: pgsql: Code review focused on new node types added by partitioning supp