About

FutoIn OptiHelp is Node.js module optimization helper.

Unlike simple benchmark tools, FutoIn OptiHelp stores base and best results. It helps you to understand how new optimization and other changes affect performance of your project over time.

Documentation --> FutoIn Guide

Author: Andrey Galkin

Installation for Node.js

Command line:

$ npm install @futoin/optihelp --save

or:

$ yarn add @futoin/optihelp --save

Concept & notes

1. Tests are unique per CPU model and Node.js version.
   • The combination is obfuscated with SHA-256 hash for minor security and ease of result management.
2. History of test results is stored under test/results in project root by default.
   • See OptiHelper dst_root option.
3. Upon first run, only benchmark results are shown. This results are called "base".
4. Upon any subsequent run, OptiHelper also shows different to "base" and "best" ever values.
5. Benchmark sequence:
   • calibration run (single shot),
   • warmup run based on test_time*warmup_ratio options,
   • actual benchmark for test_time option,
   • after all tests are done, additional benchmark passes may run (pass option).
6. Optional profiling uses v8-profiler module.
7. The result is dumped in stdout, but an overall machine readable report is also generated.
8. process.hrtime() with nanosecond resolution is used.

Usage

1. Create optihelp.js in your tests folder with Suite of tests.
2. Run the first ever cycle to get "base" results.
3. Commit changes and the results.
4. Create a branch of your project, for example optimization-refdata.
   • It would be helpful when you'll want to add extra tests.
5. Make optimization and/or other changes
6. Run the optihelp.js test again and observe the results.

Examples

const optihelp = require('@futoin/optihelp');

optihelp('Suite Name', { test_time : 5, pass : 3 } )
    .test( 'Async Test', (done) => { /* ... */; done() } )
    .test( 'Sync Test', () => { /* ... */; } )
    .start((report) => console.log(report));

API documentation

OptiHelper

Module Optimization Helper

Kind: global class

• OptiHelper
  • new OptiHelper(name, options)
  • .test(name, cb) ⇒ OptiHelper
  • .start(cb)

new OptiHelper(name, options)

C-tor

| Param | Type | Default | Description | | --- | --- | --- | --- | | name | string | | suite name | | options | object | | various options | | options.pass | integer | 2 | how many passes to run | | options.dst_root | string | "'test/results'" | result history folder | | options.test_time | float | 5 | how long to run a single pass of each test in seconds | | options.warmup_ratio | float | 1 | how long to warmup based on test_time | | options.profile_ratio | float | 0.1 | how long to profile based on test_time | | options.bench_delay | float | 1 | benchmark delays after warmup | | options.do_profile | boolean | false | run v8-profiler, if true | | options.check_prod | boolean | true | ensure running in production env | | options.report_file | string | null | store report in file, if true |

optiHelper.test(name, cb) ⇒ OptiHelper

Execute test several times

Kind: instance method of OptiHelper
Returns: OptiHelper - self for chaining

| Param | Type | Description | | --- | --- | --- | | name | string | test name | | cb | callable | test callback |

optiHelper.start(cb)

Start execution of tests

Kind: instance method of OptiHelper

| Param | Type | Description | | --- | --- | --- | | cb | callable | completion callback |