Resolver perf: Implement TreeFS.hierarchicalLookup for getClosestPackage (2/n) #1287
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
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D58062988 |
|
This pull request was exported from Phabricator. Differential Revision: D58062988 |
robhogan
force-pushed
the
export-D58062988
branch
from
04c9f67 to
c3c561f
Compare
robhogan
added a commit
to robhogan/metro
that referenced
this pull request
Aug 6, 2024
…age (2/n) (facebook#1287) Summary: Pull Request resolved: facebook#1287 Implement a new method on `TreeFS` that is able to "natively" perform hierarchical lookup, such as frequently used during node_modules resolution, config path resolution, etc., and use it initially to find the closest package scope of a given candidate path (`context.getClosestPackage()`). This operation currently dominates resolution time, with an algorithm that uses repeated `path.dirname()` and `fileSystem.lookup()`. The current algorithm is O(n^2) on path segments (~length), because the outer loop is O(n) and each lookup is O(n) - additionally, it's slow in constant time because each `path.dirname()` and `path.join()` involves unnecessarily parsing and normalising their own outputs - `path.join()` calls alone account for >30% of resolution time. The new implementation: - Performs a lookup on the start path, collecting a list of all nodes along the way. - Looks back through each node, and checks for the existence of the subpath on each one. *Benchmarks based on collecting and running all resolutions performed during a build of rn-tester:* ## Before - Total resolution time: 1210ms - Time in `getClosestPackage`: 910ms ## After - Total resolution time: 390ms (~3x faster) - Time in `getClosestPackage`: 105ms (~9x faster) Reviewed By: motiz88 Differential Revision: D58062988
|
This pull request was exported from Phabricator. Differential Revision: D58062988 |
robhogan
force-pushed
the
export-D58062988
branch
from
c3c561f to
effdb3f
Compare
robhogan
added a commit
to robhogan/metro
that referenced
this pull request
Aug 12, 2024
…age (2/n) (facebook#1287) Summary: Pull Request resolved: facebook#1287 Implement a new method on `TreeFS` that is able to "natively" perform hierarchical lookup, such as frequently used during node_modules resolution, config path resolution, etc., and use it initially to find the closest package scope of a given candidate path (`context.getClosestPackage()`). This operation currently dominates resolution time, with an algorithm that uses repeated `path.dirname()` and `fileSystem.lookup()`. The current algorithm is O(n^2) on path segments (~length), because the outer loop is O(n) and each lookup is O(n) - additionally, it's slow in constant time because each `path.dirname()` and `path.join()` involves unnecessarily parsing and normalising their own outputs - `path.join()` calls alone account for >30% of resolution time. The new implementation: - Performs a lookup on the start path, collecting a list of all nodes along the way. - Looks back through each node, and checks for the existence of the subpath on each one. *Benchmarks based on collecting and running all resolutions performed during a build of rn-tester:* ## Before - Total resolution time: 1210ms - Time in `getClosestPackage`: 910ms ## After - Total resolution time: 390ms (~3x faster) - Time in `getClosestPackage`: 105ms (~9x faster) Reviewed By: motiz88 Differential Revision: D58062988
…age (2/n) (facebook#1287) Summary: Pull Request resolved: facebook#1287 Implement a new method on `TreeFS` that is able to "natively" perform hierarchical lookup, such as frequently used during node_modules resolution, config path resolution, etc., and use it initially to find the closest package scope of a given candidate path (`context.getClosestPackage()`). This operation currently dominates resolution time, with an algorithm that uses repeated `path.dirname()` and `fileSystem.lookup()`. The current algorithm is O(n^2) on path segments (~length), because the outer loop is O(n) and each lookup is O(n) - additionally, it's slow in constant time because each `path.dirname()` and `path.join()` involves unnecessarily parsing and normalising their own outputs - `path.join()` calls alone account for >30% of resolution time. The new implementation: - Performs a lookup on the start path, collecting a list of all nodes along the way. - Looks back through each node, and checks for the existence of the subpath on each one. *Benchmarks based on collecting and running all resolutions performed during a build of rn-tester:* ## Before - Total resolution time: 1210ms - Time in `getClosestPackage`: 910ms ## After - Total resolution time: 390ms (~3x faster) - Time in `getClosestPackage`: 105ms (~9x faster) Reviewed By: motiz88 Differential Revision: D58062988
|
This pull request was exported from Phabricator. Differential Revision: D58062988 |
robhogan
force-pushed
the
export-D58062988
branch
from
effdb3f to
576f6b9
Compare
|
This pull request has been merged in 62feafa. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Summary:
Implement a new method on
TreeFSthat is able to "natively" perform hierarchical lookup, such as frequently used during node_modules resolution, config path resolution, etc., and use it initially to find the closest package scope of a given candidate path (context.getClosestPackage()).This operation currently dominates resolution time, with an algorithm that uses repeated
path.dirname()andfileSystem.lookup(). The current algorithm is O(n^2) on path segments (~length), because the outer loop is O(n) and each lookup is O(n) - additionally, it's slow in constant time because eachpath.dirname()andpath.join()involves unnecessarily parsing and normalising their own outputs -path.join()calls alone account for >30% of resolution time.The new implementation:
Benchmarks based on collecting and running all resolutions performed during a build of rn-tester:
Before
getClosestPackage: 910msAfter
getClosestPackage: 105ms (~9x faster)Differential Revision: D58062988