@agoric/polycrc
v0.1.3
Published
Fast Javascript calculation of custom CRC checksum. Prebuilt models for CRC-6, CRC-8, CRC-10, CRC-16, CRC-24, CRC-32, CRC-32C.
Downloads
36
Readme
This module provides fast calculation of CRC (Cyclic redundancy check) with custom parameters:
- width - any number 1 ... 32
- polynom
- XOR in
- XOR out
- reflect
Calculation is table-based. Tables are dynamically generated once per CRC model. There are prebuilt models for CRC-6, CRC-8, CRC-10, CRC-16, CRC-24, CRC-32, CRC-32C.
Motivation for this module were checksums (CRC-6 and CRC-10) used in IuUP (3GPP TS 25.415). It seems to be no other universal and fast CRC module in npm registry.
CRC mathematics are mostly ported from https://pycrc.org/
Installation
npm install polycrc
API
Where available, this library supports:
- Standard ECMAScript TypedArrays, DataView, ArrayBuffer
- Legacy Node.js Buffers
crc6(buffer)
crc8(buffer)
crc10(buffer)
crc16(buffer)
crc24(buffer)
crc32(buffer)
crc32c(buffer)
Calculates checksum. Returns number.
crc(width, poly, xor_in, xor_out, reflect)
Generates new CRC function with custom parameters. Returns function.
new CRC(width, poly, xor_in, xor_out, reflect)
Generates new CRC model with custom parameters. Returns class. Mostly for internal purposes.
For creating new CRC checksum functions use generator function crc().
models
module.exports.models
object holds prebuilt classes,
which can be used for pretty-printing generated CRC tables:
console.log(polycrc.models.crc8.print_table())
Example
let crypto = require('crypto')
let polycrc = require('polycrc')
let crc10 = polycrc.crc10
let result = crc10(crypto.randomBytes(1024))
console.log(result)
let crc5 = polycrc.crc(5, 0x5, 0x1f, 0x1f, true)
console.log(crc5('Input is usually a buffer, but string is ok'))
console.log(polycrc.models.crc8.print_table())
Prebuilt models
const models = {
crc1: new CRC(1, 0x01, 0x00, 0x00, false),
crc6: new CRC(6, 0x2F, 0x00, 0x00, false),
crc8: new CRC(8, 0x07, 0x00, 0x00, false),
crc10: new CRC(10, 0x233, 0x0000, 0x0000, false),
crc16: new CRC(16, 0x8005, 0x0000, 0x0000, true),
crc24: new CRC(24, 0x864CFB, 0xB704CE, 0x000000, false),
crc32: new CRC(32, 0x04C11DB7, 0xFFFFFFFF, 0xFFFFFFFF, true),
crc32c: new CRC(32, 0x1EDC6F41, 0xFFFFFFFF, 0xFFFFFFFF, true)
}
Module comparison
Here are benchmarks for several npm modules, compared to polycrc. Benchmarking tool located in examples dir.
Binary modules marked *
100 Mb of data in 300 byte chunks
algorythm module value calc/sec
crc1 polycrc 1 742093
crc1 crc 172 1588754
crc6 polycrc 8 776724
crc8 polycrc 115 1398104
crc8 crc 115 1065628
crc8 node-crc * 115 452168
crc10 polycrc 451 771580
crc16 polycrc 18086 960236
crc16 crc 18086 1046485
crc16 node-crc * 18086 484779
crc24 polycrc 3554611 775002
crc24 crc 3554611 604716
crc32 polycrc 1459514028 976329
crc32 crc 1459514028 1149756
crc32 node-crc * 1459514028 466034
crc32 polycrc 1459514028 970905
crc32 crc 1459514028 1165086
crc32 node-crc 1459514028 458094
crc32 crc-32 1459514028 1127503
crc32 buffer-crc32 1459514028 992971
crc32 cyclic-32 1459514028 1157370
crc32c polycrc 4240566998 865163
crc32c fast-crc32c(js) 4240566998 1019026
crc32c fast-crc32c * 4240566998 2818758
crc32c sse4_crc32 * 4240566998 2410524
100 Mb of data in 100 kb chunks
algorythm module value calc/sec
crc1 polycrc 0 1572
crc1 crc 193 5919
crc6 polycrc 33 1651
crc8 polycrc 27 4231
crc8 crc 27 3250
crc8 node-crc * 27 3683
crc10 polycrc 119 1670
crc16 polycrc 17776 2976
crc16 crc 17776 3240
crc16 node-crc * 17776 3592
crc24 polycrc 14484065 2438
crc24 crc 14484065 1738
crc32 polycrc 61412246 2942
crc32 crc 61412246 3518
crc32 node-crc * 61412246 3567
crc32 polycrc 61412246 2934
crc32 crc 61412246 3592
crc32 node-crc 61412246 3631
crc32 crc-32 61412246 3580
crc32 buffer-crc32 61412246 3555
crc32 cyclic-32 61412246 3567
crc32c polycrc 587481699 3011
crc32c fast-crc32c(js) 587481699 3240
crc32c fast-crc32c * 587481699 53894
crc32c sse4_crc32 * 587481699 56888
Author
Copyright (c) 2018 Vladimir Latyshev
License: MIT