Security News
New Proposed CISA Mandate Would Require Critical Infrastructure to Report Ransom Payments Within 24 Hours
CISA has proposed a set of new rules that would require critical infrastructure to report cyber incidents and ransom payments.
A little utility for doing side-by-side benchmarks in nodejs
Weekly downloads
Readme
A little utility for doing side-by-side benchmarks in nodejs.
This is not for benchmarking your HTTP servers. Use ab for that.
npm install bench
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;
}
};
require("bench").runMain()
Then, start it up with node.
$ node my-test-script.js
It'll output the scores in processes/ms, so a higher score is always better. That is, the values are kHz, not Hz.
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.
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 ms, 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
.
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.
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.
FAQs
A little utility for doing side-by-side benchmarks in nodejs
The npm package bench receives a total of 120 weekly downloads. As such, bench popularity was classified as not popular.
We found that bench demonstrated a not healthy version release cadence and project activity because the last version was released a year ago. It has 1 open source maintainer collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Security News
CISA has proposed a set of new rules that would require critical infrastructure to report cyber incidents and ransom payments.
Security News
Redis is no longer OSS, breaking its explicit commitment to remain under the BSD 3-Clause License forever. This has angered contributors who are now working to fork the software.
Product
Socket AI now enables 'AI detected potential malware' alerts by default, ensuring users benefit from AI-powered state-of-the-art malware detection without needing to opt-in.