gpu-sort

GPU accelerated asynchronous sorting


Keywords
sort, bitonic, radix, gpu, gpgpu, compute, webgl, glsl, shader, async, promises
License
MIT
Install
npm install gpu-sort@1.0.0

Documentation

gpu-sort

GPU accelerated asynchronous sorting

Build Status Coverage Status Dependency Status npm version

Why?

WebGL seems underutilized and sorting asynchronously is really handy.

Installation

npm install gpu-sort

Example

import * as gpu from "gpu-sort";

let numbers = new Float64Array([5, 4, 3, 2, 1, 0]);

// sort in place
gpu.sort(numbers);

// sort in place asynchronously
gpu.sortAsync(numbers).then(() => console.log("sorted!"));

NodeJS support

import * as gpu from "gpu-sort";

// For NodeJS an emulation library must be provided as it doesn't support WebGL
gpu.setWebGLContext(require("gl")(1, 1));

// sort using webgl emulated context
gpu.sort(foo);

Benchmarks

https://demskie.github.io/gpu-sort

Limitations

  1. Only TypedArrays are supported.
  2. When sorting 8bit and 16bit numbers GPU acceleration is disabled (as it seems rather pointless).
  3. The maximum number of elements that can be sorted is constrained by the max texture width squared. For example 4096 ^ 2 = 16,777,216 (32bit) or 4096 ^ 2 / 2 = 8,388,608 (64bit). This can be increased in the future by up to 8 times by multiplexing the data across multiple framebuffers.