Chapter 15. Parallel Query

   Table of Contents

   15.1. How Parallel Query Works
   15.2. When Can Parallel Query Be Used?
   15.3. Parallel Plans

        15.3.1. Parallel Scans
        15.3.2. Parallel Joins
        15.3.3. Parallel Aggregation
        15.3.4. Parallel Append
        15.3.5. Parallel Plan Tips

   15.4. Parallel Safety

        15.4.1. Parallel Labeling for Functions and Aggregates

   PostgreSQL can devise query plans that can leverage multiple CPUs in
   order to answer queries faster. This feature is known as parallel
   query. Many queries cannot benefit from parallel query, either due to
   limitations of the current implementation or because there is no
   imaginable query plan that is any faster than the serial query plan.
   However, for queries that can benefit, the speedup from parallel query
   is often very significant. Many queries can run more than twice as fast
   when using parallel query, and some queries can run four times faster
   or even more. Queries that touch a large amount of data but return only
   a few rows to the user will typically benefit most. This chapter
   explains some details of how parallel query works and in which
   situations it can be used so that users who wish to make use of it can
   understand what to expect.