Press n or j to go to the next uncovered block, b, p or k for the previous block.
| 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 | 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 100x 100x 100x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 3x 52x 52x 52x 52x 52x 52x 52x 52x 52x 52x 52x 52x 52x 52x 52x 52x 52x 8x 8x 52x 44x 44x 44x 52x 52x 52x 48x 48x 52x 4x 4x 48x 48x 52x 28x 52x 24x 4x 4x 4x 24x 20x 20x 20x 20x 24x 24x 24x 57x 57x 8x 57x 49x 49x 151x 151x 151x 151x 49x 57x 57x 57x 24x 24x 24x 24x 24x 52x 20x 20x 20x 52x 4x 4x 4x 4x 24x 24x 52x 82x 82x 82x 183x 183x 183x 183x 82x 82x 82x 82x 24x 52x 3x 3x 3x 3x 3x | /**
* @license Apache-2.0
*
* Copyright (c) 2024 The Stdlib Authors.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
'use strict';
// MODULES //
var f32 = require( '@stdlib/number/float64/base/to-float32' );
var isRowMajor = require( '@stdlib/ndarray/base/assert/is-row-major' );
var sfill = require( '@stdlib/blas/ext/base/sfill' ).ndarray;
var sscal = require( '@stdlib/blas/base/sscal' ).ndarray;
// FUNCTIONS //
/**
* Tests whether a provided string indicates to transpose a matrix.
*
* @private
* @param {string} str - input string
* @returns {boolean} boolean indicating whether to transpose a matrix
*
* @example
* var bool = isTransposed( 'transpose' );
* // returns true
*
* @example
* var bool = isTransposed( 'conjugate-transpose' );
* // returns true
*
* @example
* var bool = isTransposed( 'no-transpose' );
* // returns false
*/
function isTransposed( str ) { // TODO: consider moving to a separate helper utility package
return ( str !== 'no-transpose' );
}
// MAIN //
/**
* Performs one of the matrix-vector operations `y = α*A*x + β*y` or `y = α*A^T*x + β*y`, where `α` and `β` are scalars, `x` and `y` are vectors, and `A` is an `M` by `N` matrix.
*
* @private
* @param {string} trans - specifies whether `A` should be transposed, conjugate-transposed, or not transposed
* @param {NonNegativeInteger} M - number of rows in the matrix `A`
* @param {NonNegativeInteger} N - number of columns in the matrix `A`
* @param {number} alpha - scalar constant
* @param {Float32Array} A - input matrix
* @param {integer} strideA1 - stride of the first dimension of `A`
* @param {integer} strideA2 - stride of the second dimension of `A`
* @param {NonNegativeInteger} offsetA - starting index for `A`
* @param {Float32Array} x - first input vector
* @param {integer} strideX - `x` stride length
* @param {NonNegativeInteger} offsetX - starting index for `x`
* @param {number} beta - scalar constant
* @param {Float32Array} y - second input vector
* @param {integer} strideY - `y` stride length
* @param {NonNegativeInteger} offsetY - starting index for `y`
* @returns {Float32Array} `y`
*
* @example
* var Float32Array = require( '@stdlib/array/float32' );
*
* var A = new Float32Array( [ 1.0, 2.0, 3.0, 4.0, 5.0, 6.0 ] );
* var x = new Float32Array( [ 1.0, 1.0, 1.0 ] );
* var y = new Float32Array( [ 1.0, 1.0 ] );
*
* sgemv( 'no-transpose', 2, 3, 1.0, A, 3, 1, 0, x, 1, 0, 1.0, y, 1, 0 );
* // y => <Float32Array>[ 7.0, 16.0 ]
*/
function sgemv( trans, M, N, alpha, A, strideA1, strideA2, offsetA, x, strideX, offsetX, beta, y, strideY, offsetY ) { // eslint-disable-line max-params, max-len
var isrm;
var xlen;
var ylen;
var tmp;
var da0;
var da1;
var ix;
var iy;
var ia;
var i1;
var i0;
// Note on variable naming convention: da#, i# where # corresponds to the loop number, with `0` being the innermost loop...
isrm = isRowMajor( [ strideA1, strideA2 ] );
if ( isTransposed( trans ) ) {
xlen = M;
ylen = N;
} else {
xlen = N;
ylen = M;
}
// y = beta*y
if ( beta === 0.0 ) {
sfill( ylen, 0.0, y, strideY, offsetY );
} else if ( beta !== 1.0 ) {
sscal( ylen, beta, y, strideY, offsetY );
}
if ( alpha === 0.0 ) {
return y;
}
// Form: y = α*A*x + y
if (
( !isrm && !isTransposed( trans ) ) ||
( isrm && isTransposed( trans ) )
) {
if ( isrm ) {
// For row-major matrices, the last dimension has the fastest changing index...
da0 = strideA2; // offset increment for innermost loop
da1 = strideA1 - ( ylen*strideA2 ); // offset increment for outermost loop
} else { // isColMajor
// For column-major matrices, the first dimension has the fastest changing index...
da0 = strideA1; // offset increment for innermost loop
da1 = strideA2 - ( ylen*strideA1 ); // offset increment for outermost loop
}
ia = offsetA;
ix = offsetX;
for ( i1 = 0; i1 < xlen; i1++ ) {
tmp = f32( alpha * x[ ix ] );
if ( tmp === 0.0 ) {
ia += da0 * ylen;
} else {
iy = offsetY;
for ( i0 = 0; i0 < ylen; i0++ ) {
y[ iy ] += f32( A[ ia ] * tmp );
iy += strideY;
ia += da0;
}
}
ix += strideX;
ia += da1;
}
return y;
}
// Form: y = α*A^T*x + y
// ( !isrm && isTransposed( trans ) ) || ( isrm && !isTransposed( trans ) )
if ( isrm ) {
// For row-major matrices, the last dimension has the fastest changing index...
da0 = strideA2; // offset increment for innermost loop
da1 = strideA1 - ( xlen*strideA2 ); // offset increment for outermost loop
} else { // isColMajor
// For column-major matrices, the first dimension has the fastest changing index...
da0 = strideA1; // offset increment for innermost loop
da1 = strideA2 - ( xlen*strideA1 ); // offset increment for outermost loop
}
ia = offsetA;
iy = offsetY;
for ( i1 = 0; i1 < ylen; i1++ ) {
tmp = 0.0;
ix = offsetX;
for ( i0 = 0; i0 < xlen; i0++ ) {
tmp += f32( A[ ia ] * x[ ix ] );
ix += strideX;
ia += da0;
}
y[ iy ] += f32( alpha * tmp );
iy += strideY;
ia += da1;
}
return y;
}
// EXPORTS //
module.exports = sgemv;
|