kernel

About stdlib...

We believe in a future in which the web is a preferred environment for numerical computation. To help realize this future, we've built stdlib. stdlib is a standard library, with an emphasis on numerical and scientific computation, written in JavaScript (and C) for execution in browsers and in Node.js.

The library is fully decomposable, being architected in such a way that you can swap out and mix and match APIs and functionality to cater to your exact preferences and use cases.

When you use stdlib, you can be absolutely certain that you are using the most thorough, rigorous, well-written, studied, documented, tested, measured, and high-quality code out there.

To join us in bringing numerical computing to the web, get started by checking us out on GitHub, and please consider financially supporting stdlib. We greatly appreciate your continued support!

kernel

Return a kernel for applying a one-dimensional strided array function to three input ndarrays and assigning results to an output ndarray using loop blocking.

Installation

npm install @stdlib/ndarray-base-kernels-generic-ternary-strided1d-blocked

Alternatively,

To load the package in a website via a script tag without installation and bundlers, use the ES Module available on the esm branch (see README).
If you are using Deno, visit the deno branch (see README for usage intructions).
For use in Observable, or in browser/node environments, use the Universal Module Definition (UMD) build available on the umd branch (see README).

The branches.md file summarizes the available branches and displays a diagram illustrating their relationships.

To view installation and usage instructions specific to each branch build, be sure to explicitly navigate to the respective README files on each branch, as linked to above.

Usage

var kernel = require( '@stdlib/ndarray-base-kernels-generic-ternary-strided1d-blocked' );

kernel( ndims )

Returns a kernel for applying a one-dimensional strided array function to three input ndarrays and assigning results to an output ndarray using loop blocking.

var Float64Array = require( '@stdlib/array-float64' );
var ndarray2array = require( '@stdlib/ndarray-base-to-array' );
var gwhere = require( '@stdlib/blas-ext-base-ndarray-gwhere' );

// Create data buffers:
var xbuf = new Float64Array( [ 1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 1.0, 0.0 ] );
var ybuf = new Float64Array( [ 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0 ] );
var zbuf = new Float64Array( [ -1.0, -2.0, -3.0, -4.0, -5.0, -6.0, -7.0, -8.0, -9.0, -10.0, -11.0, -12.0 ] );
var wbuf = new Float64Array( [ 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 ] );

// Define the array shapes:
var xsh = [ 1, 3, 2, 2 ];
var ysh = [ 1, 3, 2, 2 ];
var zsh = [ 1, 3, 2, 2 ];
var wsh = [ 1, 3, 2, 2 ];

// Define the array strides:
var sx = [ 12, 4, 2, 1 ];
var sy = [ 12, 4, 2, 1 ];
var sz = [ 12, 4, 2, 1 ];
var sw = [ 12, 4, 2, 1 ];

// Define the index offsets:
var ox = 0;
var oy = 0;
var oz = 0;
var ow = 0;

// Create the input ndarray descriptors:
var x = {
    'dtype': 'float64',
    'data': xbuf,
    'shape': xsh,
    'strides': sx,
    'offset': ox,
    'order': 'row-major'
};

var y = {
    'dtype': 'float64',
    'data': ybuf,
    'shape': ysh,
    'strides': sy,
    'offset': oy,
    'order': 'row-major'
};

var z = {
    'dtype': 'float64',
    'data': zbuf,
    'shape': zsh,
    'strides': sz,
    'offset': oz,
    'order': 'row-major'
};

// Create an output ndarray descriptor:
var w = {
    'dtype': 'float64',
    'data': wbuf,
    'shape': wsh,
    'strides': sw,
    'offset': ow,
    'order': 'row-major'
};

// Initialize ndarray descriptors representing subarray views:
var views = [
    {
        'dtype': x.dtype,
        'data': x.data,
        'shape': [ 2, 2 ],
        'strides': [ 2, 1 ],
        'offset': x.offset,
        'order': x.order
    },
    {
        'dtype': y.dtype,
        'data': y.data,
        'shape': [ 2, 2 ],
        'strides': [ 2, 1 ],
        'offset': y.offset,
        'order': y.order
    },
    {
        'dtype': z.dtype,
        'data': z.data,
        'shape': [ 2, 2 ],
        'strides': [ 2, 1 ],
        'offset': z.offset,
        'order': z.order
    },
    {
        'dtype': w.dtype,
        'data': w.data,
        'shape': [ 2, 2 ],
        'strides': [ 2, 1 ],
        'offset': w.offset,
        'order': w.order
    }
];

// Define an input strategy:
function inputStrategy( x ) {
    return {
        'dtype': x.dtype,
        'data': x.data,
        'shape': [ 4 ],
        'strides': [ 1 ],
        'offset': x.offset,
        'order': x.order
    };
}

// Define an output strategy:
function outputStrategy( x ) {
    return x;
}

var strategy = {
    'input': inputStrategy,
    'output': outputStrategy
};

// Resolve a kernel:
var f = kernel( 2 );

// Apply strided function:
f( gwhere, [ x, y, z, w ], views, [ 1, 3 ], [ 12, 4 ], [ 12, 4 ], [ 12, 4 ], [ 12, 4 ], strategy, strategy, strategy, strategy, {} );

var arr = ndarray2array( w.data, w.shape, w.strides, w.offset, w.order );
// returns [ [ [ [ 1.0, -2.0 ], [ 3.0, -4.0 ] ], [ [ 5.0, -6.0 ], [ 7.0, -8.0 ] ], [ [ 9.0, -10.0 ], [ 11.0, -12.0 ] ] ] ]

The function accepts the following arguments:

ndims: number of loop dimensions.

If the function is provided ndims < 2 or a value greater than the maximum number of supported loop dimensions, the function returns null.

var f = kernel( 1 );
// returns null

The returned function accepts the following arguments:

fcn: function which will be applied to three one-dimensional input subarrays and should update a one-dimensional output subarray with results.
arrays: array containing three input ndarray descriptors and one output ndarray descriptor, followed by any additional ndarray arguments.
views: initialized ndarray descriptors representing subarray views.
shape: loop dimensions.
stridesX: loop dimension strides for the first input ndarray.
stridesY: loop dimension strides for the second input ndarray.
stridesZ: loop dimension strides for the third input ndarray.
stridesW: loop dimension strides for the output ndarray.
strategyX: strategy for marshaling data to and from a first input ndarray view.
strategyY: strategy for marshaling data to and from a second input ndarray view.
strategyZ: strategy for marshaling data to and from a third input ndarray view.
strategyW: strategy for marshaling data to and from an output ndarray view.
options: function options which are passed through to fcn.

kernel.kernel2d( fcn, arrays, views, shape, stridesX, stridesY, stridesZ, stridesW, strategyX, strategyY, strategyZ, strategyW, options )

Applies a one-dimensional strided array function to a list of specified dimensions in three input ndarrays and assigns results to a provided output ndarray using loop blocking.

var Float64Array = require( '@stdlib/array-float64' );
var ndarray2array = require( '@stdlib/ndarray-base-to-array' );
var gwhere = require( '@stdlib/blas-ext-base-ndarray-gwhere' );

// Create data buffers:
var xbuf = new Float64Array( [ 1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 1.0, 0.0 ] );
var ybuf = new Float64Array( [ 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0 ] );
var zbuf = new Float64Array( [ -1.0, -2.0, -3.0, -4.0, -5.0, -6.0, -7.0, -8.0, -9.0, -10.0, -11.0, -12.0 ] );
var wbuf = new Float64Array( [ 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0 ] );

// Define the array shapes:
var xsh = [ 1, 3, 2, 2 ];
var ysh = [ 1, 3, 2, 2 ];
var zsh = [ 1, 3, 2, 2 ];
var wsh = [ 1, 3, 2, 2 ];

// Define the array strides:
var sx = [ 12, 4, 2, 1 ];
var sy = [ 12, 4, 2, 1 ];
var sz = [ 12, 4, 2, 1 ];
var sw = [ 12, 4, 2, 1 ];

// Define the index offsets:
var ox = 0;
var oy = 0;
var oz = 0;
var ow = 0;

// Create the input ndarray descriptors:
var x = {
    'dtype': 'float64',
    'data': xbuf,
    'shape': xsh,
    'strides': sx,
    'offset': ox,
    'order': 'row-major'
};

var y = {
    'dtype': 'float64',
    'data': ybuf,
    'shape': ysh,
    'strides': sy,
    'offset': oy,
    'order': 'row-major'
};

var z = {
    'dtype': 'float64',
    'data': zbuf,
    'shape': zsh,
    'strides': sz,
    'offset': oz,
    'order': 'row-major'
};

// Create an output ndarray descriptor:
var w = {
    'dtype': 'float64',
    'data': wbuf,
    'shape': wsh,
    'strides': sw,
    'offset': ow,
    'order': 'row-major'
};

// Initialize ndarray descriptors representing subarray views:
var views = [
    {
        'dtype': x.dtype,
        'data': x.data,
        'shape': [ 2, 2 ],
        'strides': [ 2, 1 ],
        'offset': x.offset,
        'order': x.order
    },
    {
        'dtype': y.dtype,
        'data': y.data,
        'shape': [ 2, 2 ],
        'strides': [ 2, 1 ],
        'offset': y.offset,
        'order': y.order
    },
    {
        'dtype': z.dtype,
        'data': z.data,
        'shape': [ 2, 2 ],
        'strides': [ 2, 1 ],
        'offset': z.offset,
        'order': z.order
    },
    {
        'dtype': w.dtype,
        'data': w.data,
        'shape': [ 2, 2 ],
        'strides': [ 2, 1 ],
        'offset': w.offset,
        'order': w.order
    }
];

// Define an input strategy:
function inputStrategy( x ) {
    return {
        'dtype': x.dtype,
        'data': x.data,
        'shape': [ 4 ],
        'strides': [ 1 ],
        'offset': x.offset,
        'order': x.order
    };
}

// Define an output strategy:
function outputStrategy( x ) {
    return x;
}

var strategy = {
    'input': inputStrategy,
    'output': outputStrategy
};

// Apply strided function:
kernel.kernel2d( gwhere, [ x, y, z, w ], views, [ 1, 3 ], [ 12, 4 ], [ 12, 4 ], [ 12, 4 ], [ 12, 4 ], strategy, strategy, strategy, strategy, {} );

var arr = ndarray2array( w.data, w.shape, w.strides, w.offset, w.order );
// returns [ [ [ [ 1.0, -2.0 ], [ 3.0, -4.0 ] ], [ [ 5.0, -6.0 ], [ 7.0, -8.0 ] ], [ [ 9.0, -10.0 ], [ 11.0, -12.0 ] ] ] ]