Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
JavaScript C
tree: f0697841c7

Fetching latest commit…

Cannot retrieve the latest commit at this time

Failed to load latest commit information.
examples
lib
src
.gitignore
README.md
package.json

README.md

node bench

A little utility for doing side-by-side benchmarks in nodejs.

This is not for benchmarking your HTTP servers. Use ab for that.

Installation

$ make && sudo make install

This will create a /usr/local/bin/node-bench that you can throw your scripts into.

To uninstall, just come back here, and do:

$ sudo make uninstall

Usage

Write your script like this:

exports.compare = {
  "function wrapper" : function () {
    var x = (function (a) {
      return a;
    })("foo");
  },
  "with(){} wrapper" : function () {
    var x;
    with ({a : "foo"}) {
      x = a;
    }
  }
  "no wrapper" : function () {
    var a = "foo";
    var x = a;
  }
};

Then, you can either give it a shebang pointing at node-bench, or run node-bench against it.

$ node-bench my-test-script.js

It'll output the scores in processes/second, so a higher score is always better.

You can also export time, count, and comparecount to change the behavior slightly.

Your test script is just a plain old commonjs module, so it can include other things, update require.paths, whatever setup you need to do. Generally, it's a good idea to do this stuff in the module itself, rather than in the comparison functions, so that you can better isolate the units that you want to test.

Fields

Export the following fields from your benchmark script.

compare - The hash of functions that will be compared. The output will use the object key as the title. They're called without any arguments, in the default scope. It's assumed that you should know how to make this do whatever you need it to.

time - How long (in ms) to run the tests for. A higher value will result in more accurate tests that take longer to run. Default: 1000

compareCount - How many times to do the test runs. This should be some fairly small number. Tests are run multiple times in varying order to average out the variation due to calling one function first, a primed cache, garbage collection, etc. Higher value = more accurate, slower tests. Default: 8

countPerLap - Especially when doing asynchronous benchmarking, you may want to structure your functions so that they run a bunch of times before continuing. In these cases, to make your scores reflect the actual number of processes per second, indicate the number of runs per call in the "countPerLap" field. Default: 1

done - A function that will be called with the results of the runs when they're complete. By default, this calls a function that will analyze the results a bit and write the data to stdout.

Asynchronous Benchmarking

Just write your functions so that they take a single argument. That argument is your callback. Have fun with it.

Your callback will be fired using process.nextTick. This has a wee bit of overhead, so if you're testing something really fast, you should probably construct it to run many times before calling the callback. Check the examples/nexttick-vs-settimeout.js test for an example.

Using node-bench programmatically

Put the lib/bench.js file in your $NODE_PATH somewhere, and then do require("bench") to access its goodies. (Or put it somewhere else, and require it from there.)

WARNING!

Statistics are powerful tools, and in the wrong hands, can lead to a lot of mayhem. Please use this tool for good, and not evil.

Something went wrong with that request. Please try again.