Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

perf(semantic): reduce AstNodeId to u32 #4264

Merged
merged 1 commit into from
Jul 15, 2024

Conversation

overlookmotel
Copy link
Collaborator

@overlookmotel overlookmotel commented Jul 15, 2024

AstNodeId was a usize. This seems excessive.

Parser has a limit on size of a JS file of 4 GiB. While it is possible for a JS file of that size to create an AST with more than 1 << 32 (~4 billion) AST nodes, that would be insanely large.

So make AstNodeId u32 instead.

Copy link

graphite-app bot commented Jul 15, 2024

Your org has enabled the Graphite merge queue for merging into main

Add the label “merge” to the PR and Graphite will automatically add it to the merge queue when it’s ready to merge. Or use the label “hotfix” to add to the merge queue as a hot fix.

You must have a Graphite account and log in to Graphite in order to use the merge queue. Sign up using this link.

Copy link
Collaborator Author

overlookmotel commented Jul 15, 2024

Copy link

codspeed-hq bot commented Jul 15, 2024

CodSpeed Performance Report

Merging #4264 will improve performances by 17.64%

Comparing 07-15-perf_semantic_reduce_astnodeid_to_u32_ (8fad7db) with main (a8dc4f3)

Summary

⚡ 13 improvements
✅ 19 untouched benchmarks

Benchmarks breakdown

Benchmark main 07-15-perf_semantic_reduce_astnodeid_to_u32_ Change
minifier[antd.js] 363.8 ms 341.8 ms +6.45%
minifier[react.development.js] 4 ms 3.8 ms +6.43%
minifier[typescript.js] 672.1 ms 625.2 ms +7.5%
semantic[RadixUIAdoptionSection.jsx] 126.5 µs 107.5 µs +17.64%
semantic[antd.js] 196.3 ms 172.3 ms +13.95%
semantic[cal.com.tsx] 62.8 ms 56.7 ms +10.83%
semantic[checker.ts] 116 ms 102.7 ms +12.99%
semantic[pdf.mjs] 30 ms 26.4 ms +13.3%
transformer[RadixUIAdoptionSection.jsx] 251.2 µs 230.9 µs +8.8%
transformer[antd.js] 318.9 ms 295.6 ms +7.9%
transformer[cal.com.tsx] 80.6 ms 75.5 ms +6.85%
transformer[checker.ts] 183 ms 169.2 ms +8.14%
transformer[pdf.mjs] 49.7 ms 45.8 ms +8.37%

@overlookmotel
Copy link
Collaborator Author

Now we're talking!

bench

@overlookmotel overlookmotel marked this pull request as ready for review July 15, 2024 00:58
@overlookmotel
Copy link
Collaborator Author

@rzvxa When we were chatting a while back, you argued against making it a u32, in order to support very large files. However, I think it's hard to argue against this kind of performance boost. What do you think?

@rzvxa
Copy link
Collaborator

rzvxa commented Jul 15, 2024

@rzvxa When we were chatting a while back, you argued against making it a u32, in order to support very large files. However, I think it's hard to argue against this kind of performance boost. What do you think?

Yep, We can add a feature later on for large-file-support that would use usize. Since it is most useful for big corporate consumers and they can afford to create their own artifacts/builds.

Copy link
Collaborator

@rzvxa rzvxa left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It LGTM; it is up to you @Boshen!

@Boshen Boshen added the 0-merge Merge with Graphite Merge Queue label Jul 15, 2024
Copy link

graphite-app bot commented Jul 15, 2024

Merge activity

  • Jul 14, 10:38 PM EDT: The merge label 'merge' was detected. This PR will be added to the Graphite merge queue once it meets the requirements.
  • Jul 15, 12:31 AM EDT: Boshen added this pull request to the Graphite merge queue.
  • Jul 15, 12:43 AM EDT: Boshen merged this pull request with the Graphite merge queue.

`AstNodeId` was a `usize`. This seems excessive.

Parser has a limit on size of a JS file of 4 GiB. While it is *possible* for a JS file of that size to create an AST with more than `1 << 32` (~4 billion) AST nodes, that would be insanely large.

So make `AstNodeId` `u32` instead.
@Boshen Boshen force-pushed the 07-15-perf_semantic_do_not_record_ast_nodes_for_cfg_if_cfg_disabled branch from 8faa5e8 to 23743db Compare July 15, 2024 04:32
@Boshen Boshen force-pushed the 07-15-perf_semantic_reduce_astnodeid_to_u32_ branch from 8a26b26 to 8fad7db Compare July 15, 2024 04:33
@Boshen Boshen changed the base branch from 07-15-perf_semantic_do_not_record_ast_nodes_for_cfg_if_cfg_disabled to main July 15, 2024 04:43
@graphite-app graphite-app bot merged commit 8fad7db into main Jul 15, 2024
28 checks passed
@graphite-app graphite-app bot deleted the 07-15-perf_semantic_reduce_astnodeid_to_u32_ branch July 15, 2024 04:43
This was referenced Jul 17, 2024
Dunqing pushed a commit that referenced this pull request Jul 18, 2024
## [0.21.0] - 2024-07-18

- d7ab0b8 semantic: [**BREAKING**] Simplify node creation (#4226)
(lucab)

### Features

- af4dc01 ast: Align ts ast scope with typescript (#4253) (Dunqing)
- 83c2c62 codegen: Add option for choosing quotes; remove slow
`choose_quot` method (#4219) (Boshen)
- 5d17675 mangler: Add debug mode (#4314) (Boshen)
- e3e663b mangler: Initialize crate and integrate into minifier (#4197)
(Boshen)
- c818472 minifier: Dce conditional expression `&&` or `||` (#4190)
(Boshen)
- 8a190eb oxc: Export `oxc_mangler` (Boshen)
- 20cdb1f semantic: Align class scope with typescript (#4195) (Dunqing)
- 92ee774 semantic: Add `ScopeFlags::CatchClause` for use in CatchClause
(#4205) (Dunqing)
- 205c259 sourcemap: Support SourceMapBuilder#token_chunks (#4220)
(underfin)
- 7eb960d transformer: Decode xml character entity `&#xhhhh` and
`&#nnnn;` (#4235) (Boshen)

### Bug Fixes

- bf3d8d3 codegen: Print annotation comment inside parens for new and
call expressions (#4290) (Boshen)
- 084ab76 codegen: Use `ryu-js` for f64 to string (Boshen)
- e167ef7 codegen: Print parenthesis properly (#4245) (Boshen)
- c65198f codegen: Choose the right quote for jsx attribute string
(#4236) (Boshen)
- be82c28 codegen: Print `JSXAttributeValue::StringLiteral` directly
(#4231) (Boshen)
- 3df9e69 mangler: No shorthand `BindingProperty`; handle var hoisting
and export variables (#4319) (Boshen)
- f144082 minifier: RemoveDeadCode should visit nested expression
(#4268) (underfin)
- 66b455a oxc_codegen: Avoid print same pure comments multiple time
(#4230) (IWANABETHATGUY)
- 9a87e41 parser: Avoid crashing on invalid const modifier (#4267)
(lucab)
- 641a78b parser: Fix tests for number parsing (#4254) (overlookmotel)
- 9badac0 semantic: Avoid var hosting insert the var variable to the
`CatchClause` scope (#4337) (Dunqing)
- 95e15b6 semantic: Incorrect resolve references for `ExportSpecifier`
(#4320) (Dunqing)
- c362bf7 semantic: Incorrect resolve references for
`TSInterfaceHeritage` (#4311) (Dunqing)
- 351ecf2 semantic: Incorrect resolve references for `TSTypeQuery`
(#4310) (Dunqing)
- 1108f2a semantic: Resolve references to the incorrect symbol (#4280)
(Dunqing)
- 22d56bd semantic: Do not resolve references after `FormalParameters`
in TS type (#4241) (overlookmotel)- 1c117eb Avoid print extra semicolon
after accessor property (#4199) (IWANABETHATGUY)

### Performance

- a8dc4f3 parser: Speed up parsing numbers with `_` separators (#4259)
(overlookmotel)
- b94540d parser: Speed up parsing octal literals (#4258)
(overlookmotel)
- a7b328c parser: Faster parsing decimal numbers (#4257) (overlookmotel)
- f9d3f2e semantic: Inline ast record functions (#4272) (overlookmotel)
- 8fad7db semantic: Reduce `AstNodeId` to `u32` (#4264) (overlookmotel)
- 23743db semantic: Do not record ast nodes for cfg if cfg disabled
(#4263) (overlookmotel)
- da69076 semantic: Reduce overhead of cfg recording ast nodes (#4262)
(overlookmotel)
- cb15303 semantic: Reduce memory copies (#4216) (overlookmotel)
- ef4c1f4 semantic: Reduce lookups (#4214) (overlookmotel)
- f23e54f semantic: Recycle unresolved references hash maps (#4213)
(overlookmotel)
- 2602ce2 semantic: Reuse existing map of unresolved refs (#4206)
(lucab)

### Refactor

- 2c7bb9f ast: Pass final `ScopeFlags` into `visit_function` (#4283)
(overlookmotel)
- 3e099fe ast: Move `enter_scope` after `visit_binding_identifier`
(#4246) (Dunqing)
- aab7aaa ast/visit: Fire node events as the outermost one. (#4203)
(rzvxa)
- d1c4be0 codegen: Clean up annotation_comment (Boshen)
- 06197b8 codegen: Separate tests (Boshen)
- aa22073 codegen: Improve print API (#4196) (Boshen)
- c5731a5 semantic: Remove defunct code setting ScopeFlags twice (#4286)
(overlookmotel)
- 16698bc semantic: Move function/class-specific code into specific
visitors (#4278) (overlookmotel)
- ee16668 semantic: Rename function param (#4277) (overlookmotel)
- 25f0771 semantic: Alter syntax of `control_flow!` macro (#4275)
(overlookmotel)
- 639fd48 semantic: Comment why extra CFG enabled check (#4274)
(overlookmotel)
- c418bf5 semantic: Directly record `current_node_id` when adding a
scope (#4265) (Dunqing)
- ace4f1f semantic: Update the order of `visit_function` and `Visit`
fields in the builder to be consistent (#4248) (Dunqing)
- 8bfeabf semantic: Simplify adding `SymbolFlags::Export` (#4249)
(Dunqing)
- dc2b3c4 semantic: Add strict mode in scope flags for class definitions
(#4156) (Dunqing)
- 81ed588 semantic: Convert scope fields to IndexVecs (#4208) (lucab)
- bbe5ded semantic: Set `current_scope_id` to `scope_id` in
`enter_scope` (#4193) (Dunqing)
- 7f1addd semantic: Correct scope in CatchClause (#4192) (Dunqing)
- fc0b17d syntax: Turn the `AstNodeId::dummy` into a constant field.
(#4308) (rzvxa)
- a197e01 transformer/typescript: Remove unnecessary code (#4321)
(Dunqing)
- 1458d81 visit: Add `#[inline]` to empty functions (#4330)
(overlookmotel)

Co-authored-by: Boshen <Boshen@users.noreply.github.com>
@github-actions github-actions bot mentioned this pull request Jul 18, 2024
Dunqing pushed a commit that referenced this pull request Jul 18, 2024
## [0.21.0] - 2024-07-18

- d7ab0b8 semantic: [**BREAKING**] Simplify node creation (#4226)
(lucab)

### Features

- af4dc01 ast: Align ts ast scope with typescript (#4253) (Dunqing)
- 83c2c62 codegen: Add option for choosing quotes; remove slow
`choose_quot` method (#4219) (Boshen)
- 5d17675 mangler: Add debug mode (#4314) (Boshen)
- e3e663b mangler: Initialize crate and integrate into minifier (#4197)
(Boshen)
- c818472 minifier: Dce conditional expression `&&` or `||` (#4190)
(Boshen)
- 8a190eb oxc: Export `oxc_mangler` (Boshen)
- 20cdb1f semantic: Align class scope with typescript (#4195) (Dunqing)
- 92ee774 semantic: Add `ScopeFlags::CatchClause` for use in CatchClause
(#4205) (Dunqing)
- 205c259 sourcemap: Support SourceMapBuilder#token_chunks (#4220)
(underfin)
- 7eb960d transformer: Decode xml character entity `&#xhhhh` and
`&#nnnn;` (#4235) (Boshen)

### Bug Fixes

- bf3d8d3 codegen: Print annotation comment inside parens for new and
call expressions (#4290) (Boshen)
- 084ab76 codegen: Use `ryu-js` for f64 to string (Boshen)
- e167ef7 codegen: Print parenthesis properly (#4245) (Boshen)
- c65198f codegen: Choose the right quote for jsx attribute string
(#4236) (Boshen)
- be82c28 codegen: Print `JSXAttributeValue::StringLiteral` directly
(#4231) (Boshen)
- 3df9e69 mangler: No shorthand `BindingProperty`; handle var hoisting
and export variables (#4319) (Boshen)
- f144082 minifier: RemoveDeadCode should visit nested expression
(#4268) (underfin)
- 66b455a oxc_codegen: Avoid print same pure comments multiple time
(#4230) (IWANABETHATGUY)
- 9a87e41 parser: Avoid crashing on invalid const modifier (#4267)
(lucab)
- 641a78b parser: Fix tests for number parsing (#4254) (overlookmotel)
- 9badac0 semantic: Avoid var hosting insert the var variable to the
`CatchClause` scope (#4337) (Dunqing)
- 95e15b6 semantic: Incorrect resolve references for `ExportSpecifier`
(#4320) (Dunqing)
- c362bf7 semantic: Incorrect resolve references for
`TSInterfaceHeritage` (#4311) (Dunqing)
- 351ecf2 semantic: Incorrect resolve references for `TSTypeQuery`
(#4310) (Dunqing)
- 1108f2a semantic: Resolve references to the incorrect symbol (#4280)
(Dunqing)
- 22d56bd semantic: Do not resolve references after `FormalParameters`
in TS type (#4241) (overlookmotel)- 1c117eb Avoid print extra semicolon
after accessor property (#4199) (IWANABETHATGUY)

### Performance

- a8dc4f3 parser: Speed up parsing numbers with `_` separators (#4259)
(overlookmotel)
- b94540d parser: Speed up parsing octal literals (#4258)
(overlookmotel)
- a7b328c parser: Faster parsing decimal numbers (#4257) (overlookmotel)
- f9d3f2e semantic: Inline ast record functions (#4272) (overlookmotel)
- 8fad7db semantic: Reduce `AstNodeId` to `u32` (#4264) (overlookmotel)
- 23743db semantic: Do not record ast nodes for cfg if cfg disabled
(#4263) (overlookmotel)
- da69076 semantic: Reduce overhead of cfg recording ast nodes (#4262)
(overlookmotel)
- cb15303 semantic: Reduce memory copies (#4216) (overlookmotel)
- ef4c1f4 semantic: Reduce lookups (#4214) (overlookmotel)
- f23e54f semantic: Recycle unresolved references hash maps (#4213)
(overlookmotel)
- 2602ce2 semantic: Reuse existing map of unresolved refs (#4206)
(lucab)

### Refactor

- 2c7bb9f ast: Pass final `ScopeFlags` into `visit_function` (#4283)
(overlookmotel)
- 3e099fe ast: Move `enter_scope` after `visit_binding_identifier`
(#4246) (Dunqing)
- aab7aaa ast/visit: Fire node events as the outermost one. (#4203)
(rzvxa)
- d1c4be0 codegen: Clean up annotation_comment (Boshen)
- 06197b8 codegen: Separate tests (Boshen)
- aa22073 codegen: Improve print API (#4196) (Boshen)
- c5731a5 semantic: Remove defunct code setting ScopeFlags twice (#4286)
(overlookmotel)
- 16698bc semantic: Move function/class-specific code into specific
visitors (#4278) (overlookmotel)
- ee16668 semantic: Rename function param (#4277) (overlookmotel)
- 25f0771 semantic: Alter syntax of `control_flow!` macro (#4275)
(overlookmotel)
- 639fd48 semantic: Comment why extra CFG enabled check (#4274)
(overlookmotel)
- c418bf5 semantic: Directly record `current_node_id` when adding a
scope (#4265) (Dunqing)
- ace4f1f semantic: Update the order of `visit_function` and `Visit`
fields in the builder to be consistent (#4248) (Dunqing)
- 8bfeabf semantic: Simplify adding `SymbolFlags::Export` (#4249)
(Dunqing)
- dc2b3c4 semantic: Add strict mode in scope flags for class definitions
(#4156) (Dunqing)
- 81ed588 semantic: Convert scope fields to IndexVecs (#4208) (lucab)
- bbe5ded semantic: Set `current_scope_id` to `scope_id` in
`enter_scope` (#4193) (Dunqing)
- 7f1addd semantic: Correct scope in CatchClause (#4192) (Dunqing)
- fc0b17d syntax: Turn the `AstNodeId::dummy` into a constant field.
(#4308) (rzvxa)
- a197e01 transformer/typescript: Remove unnecessary code (#4321)
(Dunqing)
- 1458d81 visit: Add `#[inline]` to empty functions (#4330)
(overlookmotel)

Co-authored-by: Boshen <Boshen@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
0-merge Merge with Graphite Merge Queue
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants