IEEE 754 half-precision floating-point ponyfill for JavaScript
See TC39 proposal or the archive of the ES Discuss Float16Array topic for details
npm install @petamoriken/float16
Note
float16 features are supported since Deno v1.43.
deno add @petamoriken/float16
Note
float16 features are supported since Bun v1.1.23.
bun add @petamoriken/float16
import {
Float16Array, isFloat16Array, isTypedArray,
getFloat16, setFloat16,
f16round,
} from "@petamoriken/float16";
Deliver a browser/float16.mjs
or browser/float16.js
file in the npm package
from your Web server with the JavaScript Content-Type
HTTP header.
<!-- Module Scripts -->
<script type="module">
import {
Float16Array, isFloat16Array, isTypedArray,
getFloat16, setFloat16,
f16round,
} from "DEST/TO/float16.mjs";
</script>
<!-- Classic Scripts -->
<script src="DEST/TO/float16.js"></script>
<script>
const {
Float16Array, isFloat16Array, isTypedArray,
getFloat16, setFloat16,
f16round,
} = float16;
</script>
Or, you can use jsDelivr CDN.
<!-- Module Scripts -->
<script type="module">
import {
Float16Array, isFloat16Array, isTypedArray,
getFloat16, setFloat16,
f16round,
} from "https://cdn.jsdelivr.net/npm/@petamoriken/float16/+esm";
</script>
<!-- Classic Scripts -->
<script src="https://cdn.jsdelivr.net/npm/@petamoriken/float16/browser/float16.min.js"></script>
<script>
const {
Float16Array, isFloat16Array, isTypedArray,
getFloat16, setFloat16,
f16round,
} = float16;
</script>
This package only requires ES2015 features and does not use
environment-dependent features (except for inspect/
), so you can use it
without any problems. It works fine with
the current officially supported versions of Node.js.
Float16Array
implemented by Proxy
and Reflect
, so IE11 is never supported
even if you use polyfills.
lib/
and browser/
directories in the npm package have JavaScript files
already transpiled, and they have been tested automatically in the following
environments:
- Node.js: Active LTS
- Firefox: last 2 versions and ESR
- Chrome: last 2 versions
- Safari: last 2 versions
Float16Array
is similar to TypedArray
such as Float32Array
(MDN).
const array = new Float16Array([1.0, 1.1, 1.2, 1.3]);
for (const value of array) {
// 1, 1.099609375, 1.2001953125, 1.2998046875
console.log(value);
}
// Float16Array(4) [ 2, 2.19921875, 2.3984375, 2.599609375 ]
array.map((value) => value * 2);
Warning
This API returns false
for ECMAScript's native Float16Array
isFloat16Array
is a utility function to check whether the value given as an
argument is an instance of Float16Array
or not.
const buffer = new ArrayBuffer(256);
// true
isFloat16Array(new Float16Array(buffer));
// false
isFloat16Array(new Float32Array(buffer));
isFloat16Array(new Uint16Array(buffer));
isFloat16Array(new DataView(buffer));
isTypedArray
is a utility function to check whether the value given as an
argument is an instance of a type of TypedArray
or not. Unlike
util.types.isTypedArray
in Node.js, this returns true
for Float16Array
.
const buffer = new ArrayBuffer(256);
// true
isTypedArray(new Float16Array(buffer));
isTypedArray(new Float32Array(buffer));
isTypedArray(new Uint16Array(buffer));
// false
isTypedArray(new DataView(buffer));
getFloat16
and setFloat16
are similar to DataView
methods such as
DataView#getFloat32
(MDN)
and DataView#setFloat32
(MDN).
declare function getFloat16(view: DataView, byteOffset: number, littleEndian?: boolean): number;
declare function setFloat16(view: DataView, byteOffset: number, value: number, littleEndian?: boolean): void;
const buffer = new ArrayBuffer(256);
const view = new DataView(buffer);
view.setUint16(0, 0x1234);
getFloat16(view, 0); // 0.0007572174072265625
// You can append methods to DataView instance
view.getFloat16 = (...args) => getFloat16(view, ...args);
view.setFloat16 = (...args) => setFloat16(view, ...args);
view.getFloat16(0); // 0.0007572174072265625
view.setFloat16(0, Math.PI, true);
view.getFloat16(0, true); // 3.140625
f16round
is similar to Math.fround
(MDN).
This function returns nearest half-precision float representation of a number.
declare function f16round(x: number): number;
Math.fround(1.337); // 1.3370000123977661
f16round(1.337); // 1.3369140625
Float16Array
has some limitations, because it is impossible to completely reproduce the behavior of TypedArray
. Be careful when checking if it is a TypedArray
or not by using ArrayBuffer.isView
, and when using Web standards such as structuredClone
and WebGL.
Built-in TypedArray
objects use "internal slots" for built-in methods. Some
limitations exist because the Proxy
object can't trap internal slots
(explanation).
This package isn't polyfill, in other words, it doesn't change native global functions and static/prototype methods.
E.g. ArrayBuffer.isView
is the butlt-in method that checks if it has the
[[ViewedArrayBuffer]]
internal slot. It returns false
for Proxy
object
such as Float16Array
instance.
ArrayBuffer.isView(new Float32Array(10)); // true
ArrayBuffer.isView(new Float16Array(10)); // false
The structured clone algorithm copies complex JavaScript objects. It is used
internally when invoking structuredClone()
, to transfer data between Web
Workers via postMessage()
, storing objects with IndexedDB, or copying objects
for other APIs
(MDN).
It can't clone Proxy
object such as Float16Array
instance, you need to
convert it to Uint16Array
or deal with ArrayBuffer
directly.
const array = new Float16Array([1.0, 1.1, 1.2]);
const cloned = structuredClone({ buffer: array.buffer });
WebGL requires Uint16Array
for buffer or texture data whose types are
gl.HALF_FLOAT
(WebGL 2) or ext.HALF_FLOAT_OES
(WebGL 1 extension). Do not
apply the Float16Array
object directly to gl.bufferData
or gl.texImage2D
etc.
// WebGL 2 example
const vertices = new Float16Array([
-0.5, -0.5, 0,
0.5, -0.5, 0,
0.5, 0.5, 0,
]);
const buffer = gl.createBuffer();
gl.bindBuffer(gl.ARRAY_BUFFER, buffer);
// wrap in Uint16Array
gl.bufferData(gl.ARRAY_BUFFER, new Uint16Array(vertices.buffer), gl.STATIC_DRAW);
gl.vertexAttribPointer(location, 3, gl.HALF_FLOAT, false, 0, 0);
gl.bindBuffer(gl.ARRAY_BUFFER, null);
gl.enableVertexAttribArray(location);
See JSDoc comments in src/Float16Array.mjs
for details. If you don't write
hacky code, you shouldn't have any problems.
Provides custom inspection for Node.js and Deno, which makes the results of console.log
more readable.
import { Float16Array } from "@petamoriken/float16";
import { customInspect } from "@petamoriken/float16/inspect";
Float16Array.prototype[Symbol.for("nodejs.util.inspect.custom")] = customInspect;
import { Float16Array } from "https://deno.land/x/float16/mod.ts";
import { customInspect } from "https://deno.land/x/float16/inspect.ts";
// deno-lint-ignore no-explicit-any
(Float16Array.prototype as any)[Symbol.for("Deno.customInspect")] = customInspect;
Manual build and test
This repository uses corepack for package manager manager. You may have to activate yarn in corepack.
corepack enable yarn
Download devDependencies.
yarn
Build lib/
, browser/
files.
yarn run build
Build docs/
files (for browser test).
yarn run docs
This repository uses corepack for package manager manager. You may have to activate yarn in corepack.
corepack enable yarn
Download devDependencies.
yarn
NODE_ENV=test yarn build:lib
yarn test
NODE_ENV=test yarn build:browser
yarn docs
Access docs/test/index.html
with browsers.
You can access current test page
(power-assert version) in
master
branch.
MIT License
This software contains productions that are distributed under
the Apache 2.0 License.
Specifically, index.d.ts
is modified from the original
TypeScript lib files.