bench-libsass.js

var sass = require('node-sass');

var elapsed_time = function(time) {
  var elapsed = time[1] / 1000000;
  console.log("libsass in ", time[0] + " s, " + elapsed.toFixed(3) + " ms");
}

var scss = '$main-font-size: 16px;';
for (var i = 0; i < 150; i++) {
  scss += 'body { font-size: $main-font-size; }';
  scss += 'h1 { font-size: $main-font-size * 2; }';
  scss += 'pre { height: 100px - 2px; }';
}

var start = process.hrtime();
var result = sass.renderSync({
  data: scss
});

var stop = process.hrtime(start);
// console.log(result.css);
elapsed_time(stop);

bench-postcss.js

var fs = require("fs");
var postcss = require("postcss");
var calc = require("postcss-calc");
var vars = require("postcss-custom-properties");

var elapsed_time = function(time) {
  var elapsed = time[1] / 1000000;
  console.log("postcss in ", time[0] + " s, " + elapsed.toFixed(3) + " ms");
}

var css = ':root { --main-font-size: 16px; }';
for (var i = 0; i < 150; i++) {
  css += 'body { font-size: var(--main-font-size); }';
  css += 'h1 { font-size: calc(var(--main-font-size) * 2); }';
  css += 'pre { height: calc(100px - 2px); }';
}

var start = process.hrtime();

var output = postcss()
  .use(vars())
  .use(calc())
  .process(css)
  .css

var stop = process.hrtime(start);
// console.log(output);
elapsed_time(stop);

results.txt

libsass in 0 s, 135.280 ms
postcss in 0 s, 283.103 ms

test-env.txt

# cat /proc/cpuinfo
vendor_id       : GenuineIntel
cpu family      : 6
model           : 28
model name      : Intel(R) Atom(TM) CPU  330   @ 1.60GHz
stepping        : 2
microcode       : 0x219
cpu MHz         : 1596.074
cache size      : 512 KB
# node -v
v0.10.30
# node-sass -v
node-sass       3.0.0-beta.7    (Wrapper)       [JavaScript]
libsass         3.2.0-beta.6    (Sass Compiler) [C/C++]
# not sure how to fetch postcss version
# according to npm install it's 4.1.6 

I updated benchmark with more real examples: nesting, variables, mixins, math: https://github.com/postcss/postcss/blob/master/benchmark/general.js

But I got different result:

PostCSS:   47 ms
Rework:    77 ms   (1.7 times slower)
libsass:   128 ms  (2.7 times slower)
Less:      157 ms  (3.4 times slower)
Stylus:    189 ms  (4.1 times slower)
Ruby Sass: 1143 ms (24.5 times slower)

But I didn’t still use Bootstrap before mixins/variables. It contains comments, at-rules, whitespaces. It is a good test for real world parsing, so we can’t miss it.

@mgreter Can you check my updated benchmark. Maybe I loose something.