Benchmark HTML Parser Libraries

A Benchmark of javascript libraries for parsing HTML (CPU/RAM)

Results

List of result

Date: 2021-01-19T03:13:42.044Z

Library	ms/file (Mean)	Module Startup	RAM MB/file (Mean)	Max	Lib Overhead	Delta
html-dom-parser.js	31.3028ms ±13.5821ms	6.76851ms	49.1mb ±4.91mb	53.3mb	0.434mb	23.7mb
html-parser.js	24.2716ms ±16.6921ms	1.57127ms	65.7mb ±12.3mb	74.3mb	0.0614mb	44.9mb
html5.js	100.166ms ±127.430ms	16.3293ms	77.5mb ±6.58mb	82.3mb	2.58mb	50.6mb
html5parser.js	2.49572ms ±2.60323ms	0.422035ms	52.9mb ±9.60mb	61.3mb	0.0451mb	26.5mb
htmlparser.js	25.9345ms ±141.275ms	1.48171ms	55.9mb ±6.57mb	59.3mb	0.201mb	31.1mb
htmlparser2-dom.js	29.8189ms ±13.3958ms	20.6422ms	50.0mb ±5.74mb	54.7mb	3.30mb	23.2mb
htmlparser2.js	26.2480ms ±11.6788ms	24.9610ms	48.3mb ±4.27mb	51.2mb	3.44mb	19.0mb
jsdom-fragment.js	105.638ms ±48.3693ms	422.421ms	108mb ±10.5mb	132mb	40.5mb	42.7mb
jsdom-node-locations.js	102.662ms ±44.9402ms	489.440ms	109mb ±10.8mb	134mb	39.3mb	42.0mb
jsdom.js	125.054ms ±53.2103ms	397.705ms	117mb ±11.0mb	141mb	36.4mb	53.3mb
libxmljs.js	4.02398ms ±2.63628ms	4.58750ms	43.3mb ±5.67mb	49.6mb	0.520mb	16.1mb
neutron-html5parser.js	3.18042ms ±1.99863ms	1.01811ms	45.8mb ±7.86mb	50.1mb	0.0696mb	22.0mb
parse5-fragment.js	35.1238ms ±12.5502ms	13.9843ms	60.7mb ±7.56mb	73.4mb	2.31mb	30.4mb
parse5.js	36.3975ms ±13.3324ms	15.6430ms	62.0mb ±8.06mb	74.4mb	2.45mb	33.2mb
sax-wasm.js	10.2434ms ±5.75753ms	13.4242ms	90.7mb ±29.3mb	135mb	7.26mb	99.0mb
sax.js	11.1351ms ±12.4832ms	2.33130ms	62.7mb ±10.4mb	67.2mb	0.352mb	38.6mb

Notes:

• Max = The maximum amount of memory seen during all the tests.
  (You should see higher values in the real world when parsing multiple files in sequence,
  normally garbage collection isn't guaranteed to happen after each parse, like here. This is the amount of ram you will typically need for parsing a single file)

• Delta = The amount of RAM being used at the end of the benchmark after a forced Garbage Colletion.
  This shows how good or bad the library is at releasing its resources.

• Lib Overhead = Memory usage just after importing the library and running the setup()
  minus the baseline memory usage before importing the library.

---

Device summary

Node: 14.15.1 V8: 8.4.371.19-node.17 NPM: 6.14.8 OS: Mac OS X macOS Catalina 10.15.7 darwin x64 19.6.0 Device: Apple Inc. MacBookPro15,1 | CPU Intel® Core™ i7-8750H 2.20GHz 6C/12T | RAM 16 GB | GPU Intel Intel UHD Graphics 630 Built-In 1536 MB / AMD Radeon Pro 555X PCIe 4096 MB

---

Running the benchmark

1. Clone the repository
2. Run:

   npm install
   npm start

Attribution

Original HTML files from @AndreasMadsen