fast-crc32c is a CRC-32C algorithm implementation for node.js, which uses hardware acceleration (via voxer/sse4_crc32 by Anand Suresh), and fallback to software implementation when hardware acceleration fails.
CRC-32C is faster then CRC-32, which could take advantage of full CPU operand register width (64bit) instead of CRC-32's 8bit.
When using hardware acceleration, CRC-32C is about 7x ~ 9x faster than software implemented CRC-32C.
Benchmark
The 3 tested implementations are:
- sse4_crc32c Hardware accelerated CRC-32C from sse4_crc32
- js_crc32c Javascript implemented CRC-32C
- js_crc32 Javascript implemented CRC-32 from buffer-crc32
$ npm run-script benchmark
> fast-crc32c@1.0.0 benchmark /Users/xiaoyi/Projects/node-fast-crc32c
> node benchmark
sse4_crc32c_hw for inputs 1024B x 2,356,905 ops/sec ±1.23% (94 runs sampled)
sse4_crc32c_hw for inputs 16611508B, avg 2027B x 171 ops/sec ±0.54% (89 runs sampled)
sse4_crc32c_sw for inputs 1024B x 987,415 ops/sec ±0.59% (98 runs sampled)
sse4_crc32c_sw for inputs 16611508B, avg 2027B x 67.53 ops/sec ±0.62% (72 runs sampled)
js_crc32c for inputs 1024B x 353,343 ops/sec ±0.62% (99 runs sampled)
js_crc32c for inputs 16611508B, avg 2027B x 13.23 ops/sec ±4.79% (26 runs sampled)
js_crc32 for inputs 1024B x 345,803 ops/sec ±0.56% (99 runs sampled)
js_crc32 for inputs 16611508B, avg 2027B x 13.20 ops/sec ±3.96% (26 runs sampled)
| Impl | 1024B | 16611508B, avg 2027B |
|---|---|---|
| sse4_crc32c_hw | 2,356,905 ops/sec | 171 ops/sec |
| sse4_crc32c_sw | 987,415 ops/sec | 67.53 ops/sec |
| js_crc32c | 353,343 ops/sec | 13.23 ops/sec |
| js_crc32 | 345,803 ops/sec | 13.20 ops/sec |
npm install fast-crc32c --savevar crc32 = require('fast-crc32c');
var result = crc32.calculate(data, initial);- data required, String|Buffer, data to digest
- initial optional, Number, initial CRC-32C digest