@stdlib/blas-ext-base-dsnansumpw 0.2.1 → 0.3.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (17) hide show
  1. package/NOTICE +1 -1
  2. package/README.md +141 -38
  3. package/dist/index.js +7 -7
  4. package/dist/index.js.map +3 -3
  5. package/docs/types/index.d.ts +7 -7
  6. package/include/stdlib/blas/ext/base/dsnansumpw.h +8 -6
  7. package/lib/dsnansumpw.js +6 -36
  8. package/lib/dsnansumpw.native.js +4 -5
  9. package/lib/index.js +2 -5
  10. package/lib/ndarray.js +28 -30
  11. package/lib/ndarray.native.js +6 -13
  12. package/manifest.json +80 -40
  13. package/package.json +14 -8
  14. package/src/addon.c +62 -0
  15. package/src/{dsnansumpw.c → main.c} +58 -48
  16. package/include.gypi +0 -53
  17. package/src/addon.cpp +0 -117
package/lib/ndarray.js CHANGED
@@ -45,21 +45,19 @@ var BLOCKSIZE = 128;
45
45
  *
46
46
  * @param {PositiveInteger} N - number of indexed elements
47
47
  * @param {Float32Array} x - input array
48
- * @param {integer} stride - stride length
49
- * @param {NonNegativeInteger} offset - starting index
48
+ * @param {integer} strideX - stride length
49
+ * @param {NonNegativeInteger} offsetX - starting index
50
50
  * @returns {number} sum
51
51
  *
52
52
  * @example
53
53
  * var Float32Array = require( '@stdlib/array-float32' );
54
- * var floor = require( '@stdlib/math-base-special-floor' );
55
54
  *
56
55
  * var x = new Float32Array( [ 2.0, 1.0, 2.0, -2.0, -2.0, 2.0, 3.0, 4.0, NaN, NaN ] );
57
- * var N = floor( x.length / 2 );
58
56
  *
59
- * var v = dsnansumpw( N, x, 2, 1 );
57
+ * var v = dsnansumpw( 5, x, 2, 1 );
60
58
  * // returns 5.0
61
59
  */
62
- function dsnansumpw( N, x, stride, offset ) {
60
+ function dsnansumpw( N, x, strideX, offsetX ) {
63
61
  var ix;
64
62
  var s0;
65
63
  var s1;
@@ -77,13 +75,13 @@ function dsnansumpw( N, x, stride, offset ) {
77
75
  if ( N <= 0 ) {
78
76
  return 0.0;
79
77
  }
80
- if ( N === 1 || stride === 0 ) {
81
- if ( isnanf( x[ offset ] ) ) {
78
+ ix = offsetX;
79
+ if ( strideX === 0 ) {
80
+ if ( isnanf( x[ ix ] ) ) {
82
81
  return 0.0;
83
82
  }
84
- return x[ offset ];
83
+ return N * x[ ix ];
85
84
  }
86
- ix = offset;
87
85
  if ( N < 8 ) {
88
86
  // Use simple summation...
89
87
  s = 0.0;
@@ -91,64 +89,64 @@ function dsnansumpw( N, x, stride, offset ) {
91
89
  if ( isnanf( x[ ix ] ) === false ) {
92
90
  s += x[ ix ];
93
91
  }
94
- ix += stride;
92
+ ix += strideX;
95
93
  }
96
94
  return s;
97
95
  }
98
96
  if ( N <= BLOCKSIZE ) {
99
97
  // Sum a block with 8 accumulators (by loop unrolling, we lower the effective blocksize to 16)...
100
98
  s0 = ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
101
- ix += stride;
99
+ ix += strideX;
102
100
  s1 = ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
103
- ix += stride;
101
+ ix += strideX;
104
102
  s2 = ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
105
- ix += stride;
103
+ ix += strideX;
106
104
  s3 = ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
107
- ix += stride;
105
+ ix += strideX;
108
106
  s4 = ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
109
- ix += stride;
107
+ ix += strideX;
110
108
  s5 = ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
111
- ix += stride;
109
+ ix += strideX;
112
110
  s6 = ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
113
- ix += stride;
111
+ ix += strideX;
114
112
  s7 = ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
115
- ix += stride;
113
+ ix += strideX;
116
114
 
117
115
  M = N % 8;
118
116
  for ( i = 8; i < N-M; i += 8 ) {
119
117
  s0 += ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
120
- ix += stride;
118
+ ix += strideX;
121
119
  s1 += ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
122
- ix += stride;
120
+ ix += strideX;
123
121
  s2 += ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
124
- ix += stride;
122
+ ix += strideX;
125
123
  s3 += ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
126
- ix += stride;
124
+ ix += strideX;
127
125
  s4 += ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
128
- ix += stride;
126
+ ix += strideX;
129
127
  s5 += ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
130
- ix += stride;
128
+ ix += strideX;
131
129
  s6 += ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
132
- ix += stride;
130
+ ix += strideX;
133
131
  s7 += ( isnanf( x[ ix ] ) ) ? 0.0 : x[ ix ];
134
- ix += stride;
132
+ ix += strideX;
135
133
  }
136
134
  // Pairwise sum the accumulators:
137
- s = ((s0+s1) + (s2+s3)) + ((s4+s5) + (s6+s7));
135
+ s = ( (s0+s1) + (s2+s3)) + ((s4+s5) + (s6+s7) );
138
136
 
139
137
  // Clean-up loop...
140
138
  for ( i; i < N; i++ ) {
141
139
  if ( isnanf( x[ ix ] ) === false ) {
142
140
  s += x[ ix ];
143
141
  }
144
- ix += stride;
142
+ ix += strideX;
145
143
  }
146
144
  return s;
147
145
  }
148
146
  // Recurse by dividing by two, but avoiding non-multiples of unroll factor...
149
147
  n = floor( N/2 );
150
148
  n -= n % 8;
151
- return dsnansumpw( n, x, stride, ix ) + dsnansumpw( N-n, x, stride, ix+(n*stride) ); // eslint-disable-line max-len
149
+ return dsnansumpw( n, x, strideX, ix ) + dsnansumpw( N-n, x, strideX, ix+(n*strideX) ); // eslint-disable-line max-len
152
150
  }
153
151
 
154
152
 
package/lib/ndarray.native.js CHANGED
@@ -20,8 +20,7 @@
20
20
 
21
21
  // MODULES //
22
22
 
23
- var Float32Array = require( '@stdlib/array-float32' );
24
- var addon = require( './dsnansumpw.native.js' );
23
+ var addon = require( './../src/addon.node' );
25
24
 
26
25
 
27
26
  // MAIN //
@@ -31,8 +30,8 @@ var addon = require( './dsnansumpw.native.js' );
31
30
  *
32
31
  * @param {PositiveInteger} N - number of indexed elements
33
32
  * @param {Float32Array} x - input array
34
- * @param {integer} stride - stride length
35
- * @param {NonNegativeInteger} offset - starting index
33
+ * @param {integer} strideX - stride length
34
+ * @param {NonNegativeInteger} offsetX - starting index
36
35
  * @returns {number} sum
37
36
  *
38
37
  * @example
@@ -40,18 +39,12 @@ var addon = require( './dsnansumpw.native.js' );
40
39
  * var floor = require( '@stdlib/math-base-special-floor' );
41
40
  *
42
41
  * var x = new Float32Array( [ 2.0, 1.0, 2.0, -2.0, -2.0, 2.0, 3.0, 4.0, NaN, NaN ] );
43
- * var N = floor( x.length / 2 );
44
42
  *
45
- * var v = dsnansumpw( N, x, 2, 1 );
43
+ * var v = dsnansumpw( 5, x, 2, 1 );
46
44
  * // returns 5.0
47
45
  */
48
- function dsnansumpw( N, x, stride, offset ) {
49
- var view;
50
- if ( stride < 0 ) {
51
- offset += (N-1) * stride;
52
- }
53
- view = new Float32Array( x.buffer, x.byteOffset+(x.BYTES_PER_ELEMENT*offset), x.length-offset ); // eslint-disable-line max-len
54
- return addon( N, view, stride );
46
+ function dsnansumpw( N, x, strideX, offsetX ) {
47
+ return addon.ndarray( N, x, strideX, offsetX );
55
48
  }
56
49
 
57
50
 
package/manifest.json CHANGED
@@ -1,42 +1,82 @@
1
1
  {
2
- "options": {},
3
- "fields": [
4
- {
5
- "field": "src",
6
- "resolve": true,
7
- "relative": true
8
- },
9
- {
10
- "field": "include",
11
- "resolve": true,
12
- "relative": true
13
- },
14
- {
15
- "field": "libraries",
16
- "resolve": false,
17
- "relative": false
18
- },
19
- {
20
- "field": "libpath",
21
- "resolve": true,
22
- "relative": false
23
- }
24
- ],
25
- "confs": [
26
- {
27
- "src": [
28
- "./src/dsnansumpw.c"
29
- ],
30
- "include": [
31
- "./include"
32
- ],
33
- "libraries": [
34
- "-lm"
35
- ],
36
- "libpath": [],
37
- "dependencies": [
38
- "@stdlib/math-base-assert-is-nanf"
39
- ]
40
- }
41
- ]
2
+ "options": {
3
+ "task": "build"
4
+ },
5
+ "fields": [
6
+ {
7
+ "field": "src",
8
+ "resolve": true,
9
+ "relative": true
10
+ },
11
+ {
12
+ "field": "include",
13
+ "resolve": true,
14
+ "relative": true
15
+ },
16
+ {
17
+ "field": "libraries",
18
+ "resolve": false,
19
+ "relative": false
20
+ },
21
+ {
22
+ "field": "libpath",
23
+ "resolve": true,
24
+ "relative": false
25
+ }
26
+ ],
27
+ "confs": [
28
+ {
29
+ "task": "build",
30
+ "src": [
31
+ "./src/main.c"
32
+ ],
33
+ "include": [
34
+ "./include"
35
+ ],
36
+ "libraries": [],
37
+ "libpath": [],
38
+ "dependencies": [
39
+ "@stdlib/napi-export",
40
+ "@stdlib/napi-argv",
41
+ "@stdlib/napi-argv-int64",
42
+ "@stdlib/napi-argv-strided-float32array",
43
+ "@stdlib/math-base-assert-is-nanf",
44
+ "@stdlib/napi-create-double",
45
+ "@stdlib/strided-base-stride2offset",
46
+ "@stdlib/blas-base-shared"
47
+ ]
48
+ },
49
+ {
50
+ "task": "benchmark",
51
+ "src": [
52
+ "./src/main.c"
53
+ ],
54
+ "include": [
55
+ "./include"
56
+ ],
57
+ "libraries": [],
58
+ "libpath": [],
59
+ "dependencies": [
60
+ "@stdlib/math-base-assert-is-nanf",
61
+ "@stdlib/strided-base-stride2offset",
62
+ "@stdlib/blas-base-shared"
63
+ ]
64
+ },
65
+ {
66
+ "task": "examples",
67
+ "src": [
68
+ "./src/main.c"
69
+ ],
70
+ "include": [
71
+ "./include"
72
+ ],
73
+ "libraries": [],
74
+ "libpath": [],
75
+ "dependencies": [
76
+ "@stdlib/math-base-assert-is-nanf",
77
+ "@stdlib/strided-base-stride2offset",
78
+ "@stdlib/blas-base-shared"
79
+ ]
80
+ }
81
+ ]
42
82
  }
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@stdlib/blas-ext-base-dsnansumpw",
3
- "version": "0.2.1",
3
+ "version": "0.3.0",
4
4
  "description": "Calculate the sum of single-precision floating-point strided array elements, ignoring NaN values, using pairwise summation with extended accumulation, and returning an extended precision result.",
5
5
  "license": "Apache-2.0",
6
6
  "author": {
@@ -34,12 +34,19 @@
34
34
  "url": "https://github.com/stdlib-js/stdlib/issues"
35
35
  },
36
36
  "dependencies": {
37
- "@stdlib/assert-is-error": "^0.2.1",
38
- "@stdlib/math-base-assert-is-nanf": "^0.2.1",
39
- "@stdlib/math-base-special-floor": "^0.2.1",
40
- "@stdlib/utils-define-nonenumerable-read-only-property": "^0.2.1",
41
- "@stdlib/utils-library-manifest": "^0.2.1",
42
- "@stdlib/utils-try-require": "^0.2.1"
37
+ "@stdlib/assert-is-error": "^0.2.2",
38
+ "@stdlib/blas-base-shared": "^0.1.0",
39
+ "@stdlib/math-base-assert-is-nanf": "^0.2.2",
40
+ "@stdlib/math-base-special-floor": "^0.2.3",
41
+ "@stdlib/napi-argv": "^0.2.2",
42
+ "@stdlib/napi-argv-int64": "^0.2.2",
43
+ "@stdlib/napi-argv-strided-float32array": "^0.2.2",
44
+ "@stdlib/napi-create-double": "^0.0.2",
45
+ "@stdlib/napi-export": "^0.3.0",
46
+ "@stdlib/strided-base-stride2offset": "^0.1.0",
47
+ "@stdlib/utils-define-nonenumerable-read-only-property": "^0.2.2",
48
+ "@stdlib/utils-library-manifest": "^0.2.3",
49
+ "@stdlib/utils-try-require": "^0.2.2"
43
50
  },
44
51
  "devDependencies": {},
45
52
  "engines": {
@@ -80,7 +87,6 @@
80
87
  "float",
81
88
  "float32array"
82
89
  ],
83
- "__stdlib__": {},
84
90
  "funding": {
85
91
  "type": "opencollective",
86
92
  "url": "https://opencollective.com/stdlib"
package/src/addon.c ADDED
@@ -0,0 +1,62 @@
1
+ /**
2
+ * @license Apache-2.0
3
+ *
4
+ * Copyright (c) 2024 The Stdlib Authors.
5
+ *
6
+ * Licensed under the Apache License, Version 2.0 (the "License");
7
+ * you may not use this file except in compliance with the License.
8
+ * You may obtain a copy of the License at
9
+ *
10
+ * http://www.apache.org/licenses/LICENSE-2.0
11
+ *
12
+ * Unless required by applicable law or agreed to in writing, software
13
+ * distributed under the License is distributed on an "AS IS" BASIS,
14
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
15
+ * See the License for the specific language governing permissions and
16
+ * limitations under the License.
17
+ */
18
+
19
+ #include "stdlib/blas/ext/base/dsnansumpw.h"
20
+ #include "stdlib/blas/base/shared.h"
21
+ #include "stdlib/napi/export.h"
22
+ #include "stdlib/napi/argv.h"
23
+ #include "stdlib/napi/argv_int64.h"
24
+ #include "stdlib/napi/argv_strided_float32array.h"
25
+ #include "stdlib/strided/base/stride2offset.h"
26
+ #include "stdlib/napi/create_double.h"
27
+ #include <node_api.h>
28
+
29
+ /**
30
+ * Receives JavaScript callback invocation data.
31
+ *
32
+ * @param env environment under which the function is invoked
33
+ * @param info callback data
34
+ * @return Node-API value
35
+ */
36
+ static napi_value addon( napi_env env, napi_callback_info info ) {
37
+ STDLIB_NAPI_ARGV( env, info, argv, argc, 3 );
38
+ STDLIB_NAPI_ARGV_INT64( env, N, argv, 0 );
39
+ STDLIB_NAPI_ARGV_INT64( env, strideX, argv, 2 );
40
+ STDLIB_NAPI_ARGV_STRIDED_FLOAT32ARRAY( env, X, N, strideX, argv, 1 );
41
+ STDLIB_NAPI_CREATE_DOUBLE( env, API_SUFFIX(stdlib_strided_dsnansumpw)( N, X, strideX ), v );
42
+ return v;
43
+ }
44
+
45
+ /**
46
+ * Receives JavaScript callback invocation data.
47
+ *
48
+ * @param env environment under which the function is invoked
49
+ * @param info callback data
50
+ * @return Node-API value
51
+ */
52
+ static napi_value addon_method( napi_env env, napi_callback_info info ) {
53
+ STDLIB_NAPI_ARGV( env, info, argv, argc, 4 );
54
+ STDLIB_NAPI_ARGV_INT64( env, N, argv, 0 );
55
+ STDLIB_NAPI_ARGV_INT64( env, strideX, argv, 2 );
56
+ STDLIB_NAPI_ARGV_INT64( env, offsetX, argv, 3 );
57
+ STDLIB_NAPI_ARGV_STRIDED_FLOAT32ARRAY( env, X, N, strideX, argv, 1 );
58
+ STDLIB_NAPI_CREATE_DOUBLE( env, API_SUFFIX(stdlib_strided_dsnansumpw_ndarray)( N, X, strideX, offsetX ), v );
59
+ return v;
60
+ }
61
+
62
+ STDLIB_NAPI_MODULE_EXPORT_FCN_WITH_METHOD( addon, "ndarray", addon_method );
package/src/{dsnansumpw.c → main.c} RENAMED
@@ -1,7 +1,7 @@
1
1
  /**
2
2
  * @license Apache-2.0
3
3
  *
4
- * Copyright (c) 2020 The Stdlib Authors.
4
+ * Copyright (c) 2024 The Stdlib Authors.
5
5
  *
6
6
  * Licensed under the Apache License, Version 2.0 (the "License");
7
7
  * you may not use this file except in compliance with the License.
@@ -18,7 +18,8 @@
18
18
 
19
19
  #include "stdlib/blas/ext/base/dsnansumpw.h"
20
20
  #include "stdlib/math/base/assert/is_nanf.h"
21
- #include <stdint.h>
21
+ #include "stdlib/strided/base/stride2offset.h"
22
+ #include "stdlib/blas/base/shared.h"
22
23
 
23
24
  /**
24
25
  * Computes the sum of single-precision floating-point strided array elements, ignoring `NaN` values, using pairwise summation with extended accumulation, and returning an extended precision result.
@@ -31,19 +32,39 @@
31
32
  *
32
33
  * - Higham, Nicholas J. 1993. "The Accuracy of Floating Point Summation." _SIAM Journal on Scientific Computing_ 14 (4): 783–99. doi:[10.1137/0914050](https://doi.org/10.1137/0914050).
33
34
  *
34
- * @param N number of indexed elements
35
- * @param X input array
36
- * @param stride stride length
37
- * @return output value
35
+ * @param N number of indexed elements
36
+ * @param X input array
37
+ * @param strideX stride length
38
+ * @return output value
38
39
  */
39
- double stdlib_strided_dsnansumpw( const int64_t N, const float *X, const int64_t stride ) {
40
- float *xp1;
41
- float *xp2;
40
+ double API_SUFFIX(stdlib_strided_dsnansumpw)( const CBLAS_INT N, const float *X, const CBLAS_INT strideX ) {
41
+ CBLAS_INT ox = stdlib_strided_stride2offset( N, strideX );
42
+ return API_SUFFIX( stdlib_strided_dsnansumpw_ndarray )( N, X, strideX, ox );
43
+ }
44
+
45
+ /**
46
+ * Computes the sum of single-precision floating-point strided array elements, ignoring `NaN` values, using pairwise summation with extended accumulation and alternative indexing semantics, and returning an extended precision result.
47
+ *
48
+ * ## Method
49
+ *
50
+ * - This implementation uses pairwise summation, which accrues rounding error `O(log2 N)` instead of `O(N)`. The recursion depth is also `O(log2 N)`.
51
+ *
52
+ * ## References
53
+ *
54
+ * - Higham, Nicholas J. 1993. "The Accuracy of Floating Point Summation." _SIAM Journal on Scientific Computing_ 14 (4): 783–99. doi:[10.1137/0914050](https://doi.org/10.1137/0914050).
55
+ *
56
+ * @param N number of indexed elements
57
+ * @param X input array
58
+ * @param strideX stride length
59
+ * @param offsetX starting index
60
+ * @return output value
61
+ */
62
+ double API_SUFFIX(stdlib_strided_dsnansumpw_ndarray)( const CBLAS_INT N, const float *X, const CBLAS_INT strideX, const CBLAS_INT offsetX ) {
63
+ CBLAS_INT ix;
64
+ CBLAS_INT M;
65
+ CBLAS_INT n;
66
+ CBLAS_INT i;
42
67
  double sum;
43
- int64_t ix;
44
- int64_t M;
45
- int64_t n;
46
- int64_t i;
47
68
  double s0;
48
69
  double s1;
49
70
  double s2;
@@ -56,16 +77,12 @@ double stdlib_strided_dsnansumpw( const int64_t N, const float *X, const int64_t
56
77
  if ( N <= 0 ) {
57
78
  return 0.0;
58
79
  }
59
- if ( N == 1 || stride == 0 ) {
60
- if ( stdlib_base_is_nanf( X[ 0 ] ) ) {
80
+ ix = offsetX;
81
+ if ( strideX == 0 ) {
82
+ if ( stdlib_base_is_nanf( X[ ix ] ) ) {
61
83
  return 0.0;
62
84
  }
63
- return X[ 0 ];
64
- }
65
- if ( stride < 0 ) {
66
- ix = (1-N) * stride;
67
- } else {
68
- ix = 0;
85
+ return N * X[ ix ];
69
86
  }
70
87
  if ( N < 8 ) {
71
88
  // Use simple summation...
@@ -74,7 +91,7 @@ double stdlib_strided_dsnansumpw( const int64_t N, const float *X, const int64_t
74
91
  if ( !stdlib_base_is_nanf( X[ ix ] ) ) {
75
92
  sum += (double)X[ ix ];
76
93
  }
77
- ix += stride;
94
+ ix += strideX;
78
95
  }
79
96
  return sum;
80
97
  }
@@ -82,62 +99,55 @@ double stdlib_strided_dsnansumpw( const int64_t N, const float *X, const int64_t
82
99
  if ( N <= 128 ) {
83
100
  // Sum a block with 8 accumulators (by loop unrolling, we lower the effective blocksize to 16)...
84
101
  s0 = ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
85
- ix += stride;
102
+ ix += strideX;
86
103
  s1 = ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
87
- ix += stride;
104
+ ix += strideX;
88
105
  s2 = ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
89
- ix += stride;
106
+ ix += strideX;
90
107
  s3 = ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
91
- ix += stride;
108
+ ix += strideX;
92
109
  s4 = ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
93
- ix += stride;
110
+ ix += strideX;
94
111
  s5 = ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
95
- ix += stride;
112
+ ix += strideX;
96
113
  s6 = ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
97
- ix += stride;
114
+ ix += strideX;
98
115
  s7 = ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
99
- ix += stride;
116
+ ix += strideX;
100
117
 
101
118
  M = N % 8;
102
119
  for ( i = 8; i < N-M; i += 8 ) {
103
120
  s0 += ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
104
- ix += stride;
121
+ ix += strideX;
105
122
  s1 += ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
106
- ix += stride;
123
+ ix += strideX;
107
124
  s2 += ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
108
- ix += stride;
125
+ ix += strideX;
109
126
  s3 += ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
110
- ix += stride;
127
+ ix += strideX;
111
128
  s4 += ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
112
- ix += stride;
129
+ ix += strideX;
113
130
  s5 += ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
114
- ix += stride;
131
+ ix += strideX;
115
132
  s6 += ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
116
- ix += stride;
133
+ ix += strideX;
117
134
  s7 += ( stdlib_base_is_nanf( X[ ix ] ) ) ? 0.0 : (double)X[ ix ];
118
- ix += stride;
135
+ ix += strideX;
119
136
  }
120
137
  // Pairwise sum the accumulators:
121
- sum = ((s0+s1) + (s2+s3)) + ((s4+s5) + (s6+s7));
138
+ sum = ( (s0+s1) + (s2+s3) ) + ( (s4+s5) + (s6+s7) );
122
139
 
123
140
  // Clean-up loop...
124
141
  for (; i < N; i++ ) {
125
142
  if ( !stdlib_base_is_nanf( X[ ix ] ) ) {
126
143
  sum += (double)X[ ix ];
127
144
  }
128
- ix += stride;
145
+ ix += strideX;
129
146
  }
130
147
  return sum;
131
148
  }
132
149
  // Recurse by dividing by two, but avoiding non-multiples of unroll factor...
133
150
  n = N / 2;
134
151
  n -= n % 8;
135
- if ( stride < 0 ) {
136
- xp1 = (float *)X + ( (n-N)*stride );
137
- xp2 = (float *)X;
138
- } else {
139
- xp1 = (float *)X;
140
- xp2 = (float *)X + ( n*stride );
141
- }
142
- return stdlib_strided_dsnansumpw( n, xp1, stride ) + stdlib_strided_dsnansumpw( N-n, xp2, stride );
152
+ return API_SUFFIX(stdlib_strided_dsnansumpw_ndarray)( n, X, strideX, ix ) + API_SUFFIX(stdlib_strided_dsnansumpw_ndarray)( N-n, X, strideX, ix+(n*strideX) );
143
153
  }
package/include.gypi DELETED
@@ -1,53 +0,0 @@
1
- # @license Apache-2.0
2
- #
3
- # Copyright (c) 2020 The Stdlib Authors.
4
- #
5
- # Licensed under the Apache License, Version 2.0 (the "License");
6
- # you may not use this file except in compliance with the License.
7
- # You may obtain a copy of the License at
8
- #
9
- # http://www.apache.org/licenses/LICENSE-2.0
10
- #
11
- # Unless required by applicable law or agreed to in writing, software
12
- # distributed under the License is distributed on an "AS IS" BASIS,
13
- # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
14
- # See the License for the specific language governing permissions and
15
- # limitations under the License.
16
-
17
- # A GYP include file for building a Node.js native add-on.
18
- #
19
- # Main documentation:
20
- #
21
- # [1]: https://gyp.gsrc.io/docs/InputFormatReference.md
22
- # [2]: https://gyp.gsrc.io/docs/UserDocumentation.md
23
- {
24
- # Define variables to be used throughout the configuration for all targets:
25
- 'variables': {
26
- # Source directory:
27
- 'src_dir': './src',
28
-
29
- # Include directories:
30
- 'include_dirs': [
31
- '<!@(node -e "var arr = require(\'@stdlib/utils-library-manifest\')(\'./manifest.json\',{},{\'basedir\':process.cwd(),\'paths\':\'posix\'}).include; for ( var i = 0; i < arr.length; i++ ) { console.log( arr[ i ] ); }")',
32
- ],
33
-
34
- # Add-on destination directory:
35
- 'addon_output_dir': './src',
36
-
37
- # Source files:
38
- 'src_files': [
39
- '<(src_dir)/addon.cpp',
40
- '<!@(node -e "var arr = require(\'@stdlib/utils-library-manifest\')(\'./manifest.json\',{},{\'basedir\':process.cwd(),\'paths\':\'posix\'}).src; for ( var i = 0; i < arr.length; i++ ) { console.log( arr[ i ] ); }")',
41
- ],
42
-
43
- # Library dependencies:
44
- 'libraries': [
45
- '<!@(node -e "var arr = require(\'@stdlib/utils-library-manifest\')(\'./manifest.json\',{},{\'basedir\':process.cwd(),\'paths\':\'posix\'}).libraries; for ( var i = 0; i < arr.length; i++ ) { console.log( arr[ i ] ); }")',
46
- ],
47
-
48
- # Library directories:
49
- 'library_dirs': [
50
- '<!@(node -e "var arr = require(\'@stdlib/utils-library-manifest\')(\'./manifest.json\',{},{\'basedir\':process.cwd(),\'paths\':\'posix\'}).libpath; for ( var i = 0; i < arr.length; i++ ) { console.log( arr[ i ] ); }")',
51
- ],
52
- }, # end variables
53
- }