mirror of
https://github.com/danog/dart-sass.git
synced 2025-01-22 22:02:00 +01:00
ed73c2c053
As expected, this substantially improves the JS performance in logic-heavy benchmarks. I believe the other changes are just noise.
191 lines
6.4 KiB
Markdown
191 lines
6.4 KiB
Markdown
These benchmarks are informal and only intended to give us a general sense of
|
|
the benefit Dart Sass could provide relative to other implementations.
|
|
|
|
This was tested against:
|
|
|
|
* libsass 0f3d6ad1 and sassc 4674821 compiled with g++ (Debian 7.3.0-18) 7.3.0.
|
|
* Dart Sass 50a45a7 on Dart 2.4.0 and Node v12.0.0.
|
|
* Ruby Sass 8d1edc76 on ruby 2.5.3p105 (2018-10-18 revision 65156) [x86_64-linux].
|
|
|
|
on Debian x64 with Intel Xeon E5-1650 v3 @ 3.50GHz. The Dart Sass
|
|
[application snapshot][] was trained on the `tool/app-snapshot-input.scss` file.
|
|
|
|
[application snapshot]: https://github.com/dart-lang/sdk/wiki/Snapshots
|
|
|
|
# Measurements
|
|
|
|
I ran five instances of each configuration and recorded the fastest time.
|
|
|
|
## Small Plain CSS
|
|
|
|
Running on a file containing 4 instances of `.foo {a: b}`:
|
|
|
|
* sassc: 0.005s
|
|
* Dart Sass from a script snapshot: 0.216s
|
|
* Dart Sass native executable: 0.018s
|
|
* Dart Sass on Node.js: 0.209s
|
|
* Ruby Sass with a hot cache: 0.148s
|
|
|
|
Based on these numbers, Dart Sass from a native executable is approximately:
|
|
|
|
* 3.6x slower than libsass
|
|
* 11.6x faster than Dart Sass on Node
|
|
* 8.2x faster than Ruby Sass
|
|
|
|
## Large Plain CSS
|
|
|
|
Running on a file containing 2^17 instances of `.foo {a: b}`:
|
|
|
|
* sassc: 1.705s
|
|
* Dart Sass from a script snapshot: 1.550s
|
|
* Dart Sass native executable: 1.518s
|
|
* Dart Sass on Node.js: 2.765s
|
|
* Ruby Sass with a hot cache: 10.965s
|
|
|
|
Based on these numbers, Dart Sass from a native executable is approximately:
|
|
|
|
* 1.1x faster than libsass
|
|
* 1.8x faster than Dart Sass on Node
|
|
* 7.2x faster than Ruby Sass
|
|
|
|
## Preceding Sparse `@extend`
|
|
|
|
Running on a file containing `.x {@extend .y}`, 2^17 instances of `.foo {a: b}`, and then `.y {a: b}`:
|
|
|
|
* sassc: 1.712s
|
|
* Dart Sass from a script snapshot: 1.613s
|
|
* Dart Sass native executable: 1.582s
|
|
* Dart Sass on Node.js: 2.739s
|
|
* Ruby Sass with a hot cache: 16.472s
|
|
|
|
Based on these numbers, Dart Sass from a native executable is approximately:
|
|
|
|
* 1.1x faster than libsass
|
|
* 1.7x faster than Dart Sass on Node
|
|
* 10.4x faster than Ruby Sass
|
|
|
|
## Following Sparse `@extend`
|
|
|
|
Running on a file containing `.y {a: b}`, 2^17 instances of `.foo {a: b}`, and then `.x {@extend .y}`:
|
|
|
|
* sassc: 1.701s
|
|
* Dart Sass from a script snapshot: 1.568s
|
|
* Dart Sass native executable: 1.543s
|
|
* Dart Sass on Node.js: 2.821s
|
|
* Ruby Sass with a hot cache: 16.469s
|
|
|
|
Based on these numbers, Dart Sass from a native executable is approximately:
|
|
|
|
* 1.1x faster than libsass
|
|
* 1.8x faster than Dart Sass on Node
|
|
* 10.7x faster than Ruby Sass
|
|
|
|
## Preceding Dense `@extend`
|
|
|
|
Running on a file containing `.bar {@extend .foo}` followed by 2^17 instances of `.foo {a: b}`:
|
|
|
|
* sassc: 2.260s
|
|
* Dart Sass from a script snapshot: 2.405s
|
|
* Dart Sass native executable: 2.526s
|
|
* Dart Sass on Node.js: 5.612s
|
|
* Ruby Sass with a hot cache: 28.690s
|
|
|
|
Based on these numbers, Dart Sass from a native executable is approximately:
|
|
|
|
* 1.1x slower than libsass
|
|
* 2.2x faster than Dart Sass on Node
|
|
* 11.4x faster than Ruby Sass
|
|
|
|
## Following Dense `@extend`
|
|
|
|
Running on a file containing 2^17 instances of `.foo {a: b}` followed by `.bar {@extend .foo}`:
|
|
|
|
* sassc: 2.289s
|
|
* Dart Sass from a script snapshot: 2.396s
|
|
* Dart Sass native executable: 2.457s
|
|
* Dart Sass on Node.js: 6.319s
|
|
* Ruby Sass with a hot cache: 28.708s
|
|
|
|
Based on these numbers, Dart Sass from a native executable is approximately:
|
|
|
|
* 1.1x slower than libsass
|
|
* 2.6x faster than Dart Sass on Node
|
|
* 11.7x faster than Ruby Sass
|
|
|
|
## Bootstrap
|
|
|
|
Running on a file containing 16 instances of importing the Bootstrap framework:
|
|
|
|
* sassc: 0.767s
|
|
* Dart Sass from a script snapshot: 1.534s
|
|
* Dart Sass native executable: 0.955s
|
|
* Dart Sass on Node.js: 3.156s
|
|
* Ruby Sass with a hot cache: 12.521s
|
|
|
|
Based on these numbers, Dart Sass from a native executable is approximately:
|
|
|
|
* 1.2x slower than libsass
|
|
* 3.3x faster than Dart Sass on Node
|
|
* 13.1x faster than Ruby Sass
|
|
|
|
## a11ycolor
|
|
|
|
Running on a file containing test cases for a computation-intensive color-processing library:
|
|
|
|
* sassc: 0.248s
|
|
* Dart Sass from a script snapshot: 0.736s
|
|
* Dart Sass native executable: 0.565s
|
|
* Dart Sass on Node.js: 1.043s
|
|
* Ruby Sass with a hot cache: 5.091s
|
|
|
|
Based on these numbers, Dart Sass from a native executable is approximately:
|
|
|
|
* 2.3x slower than libsass
|
|
* 1.8x faster than Dart Sass on Node
|
|
* 9.0x faster than Ruby Sass
|
|
|
|
## Susy
|
|
|
|
Running on a file containing test cases for the computation-intensive Susy grid framework:
|
|
|
|
* sassc: 0.248s
|
|
* Dart Sass from a script snapshot: 0.673s
|
|
* Dart Sass native executable: 0.237s
|
|
* Dart Sass on Node.js: 0.990s
|
|
* Ruby Sass with a hot cache: 1.527s
|
|
|
|
Based on these numbers, Dart Sass from a native executable is approximately:
|
|
|
|
* identical to libsass
|
|
* 4.2x faster than Dart Sass on Node
|
|
* 6.4x faster than Ruby Sass
|
|
|
|
# Prior Measurements
|
|
|
|
* [1.22.4](https://github.com/sass/dart-sass/blob/a7172a2b1dd48b339e5d57159ed364ffb9f5812e/perf.md).
|
|
* [1.20.2](https://github.com/sass/dart-sass/blob/4b7699291c9f69533d25980d23b0647266b665f2/perf.md).
|
|
* [1.13.4](https://github.com/sass/dart-sass/blob/b6ccc91a138e75420227ff79381c5f70e60254f1/perf.md).
|
|
* [1.6.0](https://github.com/sass/dart-sass/blob/048cbe197a77e1cf4b837a40a5acb737e949fd5c/perf.md).
|
|
* [1.0.0-alpha.8](https://github.com/sass/dart-sass/blob/be44245a849f2bb18b5ca1fc74f3043a36da17f0/perf.md).
|
|
* [Pre-alpha, 30 September 2016](https://github.com/sass/dart-sass/blob/169370bf18fd01d0618b0fc00d9db33e2fc52aa7/perf.md).
|
|
* [Pre-alpha, 19 August 2016](https://github.com/sass/dart-sass/blob/4bea13cfe57d9e3c7f1f8580b80c59abe1cfabf8/perf.md).
|
|
* [Pre-alpha, 15 July 2016](https://github.com/sass/dart-sass/blob/a3e00059c4371bfde9afada1759d8484aee05584/perf.md).
|
|
|
|
# Conclusions
|
|
|
|
This is the first measurement with Dart Sass running as ahead-of-time-compiled
|
|
native code, and the results are encouraging. It's well below the 100ms
|
|
threshold for tiny files, and it's on par with SassC for most test cases. SassC
|
|
still leads for tests with many extends, although only slightly, and for one of
|
|
our real-world test cases (although Dart Sass leads in others). The two
|
|
implementations can be fairly described as having about the same performance
|
|
overall.
|
|
|
|
Dart Sass on Node is still substantially slower than on the Dart VM, and that
|
|
relative slowdown becomes more pronounced as the raw Dart code becomes faster.
|
|
Solutions for this such as [the embedded protocol][] or [WebAssembly support][]
|
|
are becoming more and more important.
|
|
|
|
[embedded Dart Sass]: https://github.com/sass/sass-embedded-protocol
|
|
[Dart WebAssembly support]: https://github.com/dart-lang/sdk/issues/32894
|