mirror of
https://github.com/FuelLabs/sway.git
synced 2025-08-14 23:50:57 +00:00
f607a674e3
## Description This PR is a prequel to https://github.com/FuelLabs/sway/pull/7015, trying to optimize the compiler to alleviate the hit of having more items, `impl` etc... We have here two optimizations: 1 - We spend a lot of time counting newlines when converting byte offsets to `LineCol`. Now, we calculate line offsets just once, and use binary search later to find which line a byte offset is at. 2 - `QueryEngine::get_programs_cache_entry` was cloning the whole `TyProgram`. That is why `did_change_with_caching` was always getting worse, as the program was increasing in size. Now all compilation stages are behind `Arc`, which makes the `clone` free. ## Analysis Putting a `dbg!(...)` like the image below, and calling `counts` (https://github.com/nnethercote/counts). ``` cargo bench -- traverse 2>&1 | grep "sway-types/src/span.rs:29:9" | counts ```  I get the following results: ``` 972102 counts ( 1) 156720 (16.1%, 16.1%): [sway-types/src/span.rs:29:9] self.pos = 0 ( 2) 15900 ( 1.6%, 17.8%): [sway-types/src/span.rs:29:9] self.pos = 104 ( 3) 15840 ( 1.6%, 19.4%): [sway-types/src/span.rs:29:9] self.pos = 107 ( 4) 2280 ( 0.2%, 19.6%): [sway-types/src/span.rs:29:9] self.pos = 19281 ( 5) 2280 ( 0.2%, 19.9%): [sway-types/src/span.rs:29:9] self.pos = 19285 ( 6) 2280 ( 0.2%, 20.1%): [sway-types/src/span.rs:29:9] self.pos = 19287 ( 7) 2280 ( 0.2%, 20.3%): [sway-types/src/span.rs:29:9] self.pos = 19292 ( 8) 2280 ( 0.2%, 20.6%): [sway-types/src/span.rs:29:9] self.pos = 19323 ( 9) 2280 ( 0.2%, 20.8%): [sway-types/src/span.rs:29:9] self.pos = 19327 ( 10) 2280 ( 0.2%, 21.0%): [sway-types/src/span.rs:29:9] self.pos = 19329 ( 11) 2280 ( 0.2%, 21.3%): [sway-types/src/span.rs:29:9] self.pos = 19334 ( 12) 870 ( 0.1%, 21.4%): [sway-types/src/span.rs:29:9] self.pos = 4285 ... ``` This means that `line_col` is being called 972k times. 16% is for position zero, which should be trivial. The rest will iterate the whole file source code to count number of lines. Making the real complexity of the work here something like `O(qty * self.pos)`. And some values of `self.pos` are not trivial at all. ## Checklist - [ ] I have linked to any relevant issues. - [ ] I have commented my code, particularly in hard-to-understand areas. - [ ] I have updated the documentation where relevant (API docs, the reference, and the Sway book). - [ ] If my change requires substantial documentation changes, I have [requested support from the DevRel team](https://github.com/FuelLabs/devrel-requests/issues/new/choose) - [ ] I have added tests that prove my fix is effective or that my feature works. - [ ] I have added (or requested a maintainer to add) the necessary `Breaking*` or `New Feature` labels where relevant. - [ ] I have done my best to ensure that my PR adheres to [the Fuel Labs Code Review Standards](https://github.com/FuelLabs/rfcs/blob/master/text/code-standards/external-contributors.md). - [ ] I have requested a review from the relevant team or maintainers. |
||
|---|---|---|
| .. | ||
| mod.rs | ||