@marco_ciaramella/sha256-gpu
v1.1.4
Published
SHA-256 over WebGPU
Downloads
44
Maintainers
Readme
sha256-gpu
A GPU version of SHA-256 using WebGPU.
Install
npm i @marco_ciaramella/sha256-gpu
Usage
Function sha256_gpu
takes and array of messages of the same size and computes the SHA-256 for each message in parallel. A message must be in binary format as a Uint8Array
(byte array). For example if you want the SHA-256 of integer 1 you must pass it as new Uint8Array([0x01, 0x00, 0x00, 0x00])
.
import { sha256_gpu } from "@marco_ciaramella/sha256-gpu";
// each message in messages must have the same size
const messages = [
new Uint8Array([0x01, 0x00, 0x00, 0x00]), // binary representation of integer 1
new Uint8Array([0x02, 0x00, 0x00, 0x00]), // binary representation of integer 2
new Uint8Array([0x03, 0x00, 0x00, 0x00]), // binary representation of integer 3
new Uint8Array([0x04, 0x00, 0x00, 0x00]), // binary representation of integer 4
new Uint8Array([0x05, 0x00, 0x00, 0x00]), // binary representation of integer 5
new Uint8Array([0x06, 0x00, 0x00, 0x00]), // binary representation of integer 6
new Uint8Array([0x07, 0x00, 0x00, 0x00]), // binary representation of integer 7
new Uint8Array([0x08, 0x00, 0x00, 0x00]), // binary representation of integer 8
new Uint8Array([0x09, 0x00, 0x00, 0x00]) // binary representation of integer 9
];
// compute the SHA-256 in parallel
const hashes = await sha256_gpu(messages);
// print the result
for (let i = 0; i < hashes.length; i += 32) {
console.log(hashes.subarray(i, i + 32).reduce((a, b) => a + b.toString(16).padStart(2, '0'), ''));
}
Note
Use this module when you have a lot of messages and want to compute SHA-256 at once. Don't use this module when you need the SHA-256 for one message at a time. In this case the CPU version performs better.