Skip to content

wikis Search Results · repo:postgrespro/aqo language:PLpgSQL

10 results
 (171 ms)

10 results

inpostgrespro/aqo (press backspace or delete to remove)

Intro If we make 'JOB' benchmark results with specific preferences ( script ): fsync = off max_parallel_workers_per_gather = 0 statement_timeout = 600000 from_collapse_limit = 20 join_collapse_limit ...
  • Last updated
    on Apr 18, 2022

Join Order Benchmark (AQO v. 1.1) Join Order Benchmark (AQO v. 1.2)
  • Last updated
    on Feb 6, 2020

Common differences c-cpp.yml - names of branches, patch filename. stable13 - > stable14 QueryID machinery. From stable14 we can use the jumbling technique to generate queryid or another lightweight ...
  • Last updated
    on Dec 22, 2022

Structure of main branches: master - development, may be non-stable stable9_6 - branch for PostgreSQL 9.6 (no longer supported). stable11 - for PostgreSQL 10 and 11 stable12, stable13, stable14 ... - for ...
  • Last updated
    on Oct 6, 2022

How AQO accelerates queries execution over constant dataset postgresql.conf custom options: shared_preload_libraries = 'aqo' aqo.mode = 'intelligent' min_parallel_table_scan_size = 0 min_parallel_index_scan_size ...
  • Last updated
    on Feb 6, 2020

How AQO v.1.2 accelerates queries execution over constant dataset postgresql.conf custom options: shared_preload_libraries = 'aqo' min_parallel_table_scan_size = 0 min_parallel_index_scan_size ...
  • Last updated
    on Feb 6, 2020

To realize how stable AQO code is, before a release it should be tested according to the steps: make check , make check-world , of course. Stress test. See the bash script: ./configure --enable-tap-tests ...
  • Last updated
    on Jul 12, 2022

Why are we not get the optimal query plan? Planner get incorrect cost estimations for plan nodes. GEQO algorithm can't achieve global optimum. If we use AQO we have correct estimations for each node ...
  • Last updated
    on Feb 21, 2020

GUCs aqo.mode - defines how AQO works with incoming queries. Possible modes: DISABLED , CONTROLLED , LEARN , INTELLIGENT and FROZEN . aqo.show_hash - hash value computed by a query tree which have a sense ...
  • Last updated
    on Feb 6, 2023

GUCs aqo.min_neighbors_for_predicting - minimum number of neighbors is used for cardinality prediction: if there are fewer of them, AQO won't make prediction. aqo.predict_with_few_neighbors - enable ...
  • Last updated
    on Feb 7, 2023