This repository has been archived by the owner on Sep 27, 2019. It is now read-only.
Optimizer refactor and cost model additions #1484
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
linmagit
reviewed
Oct 20, 2018
| @@ -35,9 +37,16 @@ class OptimizerMetadata { | |||
| settings::SettingId::task_execution_timeout)), | |||
| timer(Timer<std::milli>()) {} | |||
There was a problem hiding this comment.
Similarity, do we still need this method here?
| namespace peloton { | ||
| namespace optimizer { | ||
|
|
||
| double PostgresCostCalculator::CalculateCost( |
There was a problem hiding this comment.
Just curious, is this exactly the same of Postgres query optimizer? Including the DEFAULT_TUPLE_COST or DEFAULT_INDEX_TUPLE_COST?
There was a problem hiding this comment.
Postgres does it a bit different as far as I can tell. While we have a cost per tuple, postgres estimates a cost for hashing a predicate to cost hash joins. This is something we could try to implement but would require more complex estimators that should probably be in a separate PR
added 12 commits
October 26, 2018 16:01
…sts, along with more optimizer test utility functions
|
@GustavoAngulo I would not worry about getting this to work on Travis since we're not maintaining it. |
added 10 commits
January 24, 2019 13:57
This reverts commit fcbf161.
GustavoAngulo
changed the title
apavlo
approved these changes
Feb 13, 2019
mtunique
pushed a commit
to mtunique/peloton
that referenced
this pull request
Apr 16, 2019
* Separate cost model from optimizer and add postgres cost model * Fix bug overflow in Analyze * Changes to cost model construction * Fix bug in stats hashing * Fix to commutativity of equality comparison expressions * Revert group equality, this works * Move Postgres cost model to header file and add some starting plan tests, along with more optimizer test utility functions * Remove printf * Remove old optimizer constructor * Fix unused variable * Testing if changing llvm path fixes travis * Did not work : ( * Ok trying changing to 3.9.1_2 * Revert "Fix bug overflow in Analyze" This reverts commit fcbf161. * Update LLVM dir in travis config * Add trivial cost model * Move files into stats folder * Add test cases for trivial cost model * Delete cost.h and cost.cpp that were commented out * Cost model name and directory refactoring * Fix three join test
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR adds a ton of different things before we move over the optimizer to terrier:
optimizer/stats, and all cost model files are underoptimizer/cost_modeloptimizer/cost_model/trivial_cost_model.h. This can be used for debugging or trivial execution. It does not use any statistics, it can be thought of as heuristic based.optimizer/stats/stats_util.h