Re: QueryParallelism

[Alessandro Fogli]

Hello, Thanks for your answer! I would like to run a TPC-H benchmark but none of the queries uses more than a single core. I used these indices:  CREATE INDEX i_n_regionkey ON nation (n_regionkey); CREATE INDEX i_s_nationkey ON supplier (s_nationkey); CREATE INDEX i_c_nationkey ON customer (c_nationkey); CREATE INDEX i_ps_suppkey ON partsupp (ps_suppkey); CREATE INDEX i_ps_partkey ON partsupp (ps_partkey); CREATE INDEX i_o_custkey ON orders (o_custkey); CREATE INDEX i_l_orderkey ON lineitem (l_orderkey); CREATE INDEX i_l_suppkey_partkey ON lineitem (l_partkey, l_suppkey); Even when I changed the query thread pool size I was unable to use more than one thread for a single query. I'm attaching you the file with which I created the tables. Thanks Best, Alessandro

dss.ddl
Description: Binary data

Il giorno 1 apr 2020, alle ore 09:13, Ilya Kasnacheev <ilya.kasnach...@gmail.com> ha scritto: Hello! Can you please specify what is the operation that you want to parallelize? Are you sure it uses index? Please note that you may also need to increase query thread pool size since its threads are used up by parallelization. Regards, -- Ilya Kasnacheev ср, 1 апр. 2020 г. в 05:22, Alessandro Fogli <alessandro.fog...@gmail.com>: Hello, I would like to know if there is a way to increase the level of parallelism within a single node. The Default value appears to be one. I already tried adding the property “<property name="queryParallelism" value="32"/>" in the xml file but it didn't work. Thanks Best regards, Alessandro