feat: add C ndarray API and refactor blas/ext/base/dsnannsumors

PR-URL: #3086

Reviewed-by: Philipp Burckhardt <[email protected]>

Author: headlessNode

lib/node_modules/@stdlib/blas/ext/base/dsnannsumors/README.md

@@ -55,11 +55,11 @@ The function has the following parameters:
 
 -   **N**: number of indexed elements.
 -   **x**: input [`Float32Array`][@stdlib/array/float32].
--   **strideX**: index increment for `x`.
+-   **strideX**: stride length for `x`.
 -   **out**: output [`Float64Array`][@stdlib/array/float64] whose first element is the sum and whose second element is the number of non-NaN elements.
--   **strideOut**: index increment for `out`.
+-   **strideOut**: stride length for `out`.
 
-The `N` and stride parameters determine which elements are accessed at runtime. For example, to compute the sum of every other element in `x`,
+The `N` and stride parameters determine which elements are accessed at runtime. For example, to compute the sum of every other element:
 
 ```javascript
 var Float32Array = require( '@stdlib/array/float32' );
@@ -92,7 +92,7 @@ var v = dsnannsumors( 4, x1, 2, out1, 1 );
 
 #### dsnannsumors.ndarray( N, x, strideX, offsetX, out, strideOut, offsetOut )
 
-Computes the sum of single-precision floating-point strided array elements, ignoring `NaN` values and using ordinary recursive summation with extended accumulation and alternative indexing semantics.
+Computes the sum of single-precision floating-point strided array elements, ignoring `NaN` values and using ordinary recursive summation with extended accumulation and alternative indexing semantics and returning an extended precision result.
 
 ```javascript
 var Float32Array = require( '@stdlib/array/float32' );
@@ -110,7 +110,7 @@ The function has the following additional parameters:
 -   **offsetX**: starting index for `x`.
 -   **offsetOut**: starting index for `out`.
 
-While [`typed array`][mdn-typed-array] views mandate a view offset based on the underlying `buffer`, the `offset` parameter supports indexing semantics based on a starting index. For example, to calculate the sum of every other value in `x` starting from the second value
+While [`typed array`][mdn-typed-array] views mandate a view offset based on the underlying buffer, the offset parameters support indexing semantics based on starting indices. For example, to calculate the sum of every other element starting from the second element:
 
 ```javascript
 var Float32Array = require( '@stdlib/array/float32' );
@@ -171,6 +171,132 @@ console.log( out );
 
 <!-- /.examples -->
 
+<!-- C interface documentation. -->
+
+* * *
+
+<section class="c">
+
+## C APIs
+
+<!-- Section to include introductory text. Make sure to keep an empty line after the intro `section` element and another before the `/section` close. -->
+
+<section class="intro">
+
+</section>
+
+<!-- /.intro -->
+
+<!-- C usage documentation. -->
+
+<section class="usage">
+
+### Usage
+
+```c
+#include "stdlib/blas/ext/base/dsnannsumors.h"
+```
+
+#### stdlib_strided_dsnannsumors( N, \*X, strideX, \*n )
+
+Computes the sum of single-precision floating-point strided array elements, ignoring `NaN` values, using ordinary recursive summation with extended accumulation, and returning an extended precision result.
+
+```c
+const float x[] = { 1.0f, -2.0f, 0.0/0.0, 2.0f };
+CBLAS_INT n = 0;
+
+double v = stdlib_strided_dsnannsumors( 4, x, 1, &n );
+// returns 1.0
+```
+
+The function accepts the following arguments:
+
+-   **N**: `[in] CBLAS_INT` number of indexed elements.
+-   **X**: `[in] float*` input array.
+-   **strideX**: `[in] CBLAS_INT` stride length for `X`.
+-   **n**: `[out] CBLAS_INT*` number of non-NaN elements.
+
+```c
+double stdlib_strided_dsnannsumors( const CBLAS_INT N, const float *X, const CBLAS_INT strideX, CBLAS_INT *n );
+```
+
+#### stdlib_strided_dsnannsumors_ndarray( N, \*X, strideX, offsetX, \*n )
+
+Computes the sum of single-precision floating-point strided array elements, ignoring `NaN` values and using ordinary recursive summation with extended accumulation and alternative indexing semantics and returning an extended precision result.
+
+```c
+const float x[] = { 1.0f, -2.0f, 0.0/0.0, 2.0f };
+CBLAS_INT n = 0;
+
+double v = stdlib_strided_dsnannsumors_ndarray( 4, x, 1, 0, &n );
+// returns 1.0
+```
+
+The function accepts the following arguments:
+
+-   **N**: `[in] CBLAS_INT` number of indexed elements.
+-   **X**: `[in] float*` input array.
+-   **strideX**: `[in] CBLAS_INT` stride length for `X`.
+-   **offsetX**: `[in] CBLAS_INT` starting index for `X`.
+-   **n**: `[out] CBLAS_INT*` number of non-NaN elements.
+
+```c
+double stdlib_strided_dsnannsumors_ndarray( const CBLAS_INT N, const float *X, const CBLAS_INT strideX, const CBLAS_INT offsetX, CBLAS_INT *n );
+```
+
+</section>
+
+<!-- /.usage -->
+
+<!-- C API usage notes. Make sure to keep an empty line after the `section` element and another before the `/section` close. -->
+
+<section class="notes">
+
+</section>
+
+<!-- /.notes -->
+
+<!-- C API usage examples. -->
+
+<section class="examples">
+
+### Examples
+
+```c
+#include "stdlib/blas/ext/base/dsnannsumors.h"
+#include "stdlib/blas/base/shared.h"
+#include <stdio.h>
+
+int main( void ) {
+    // Create a strided array:
+    const float x[] = { 1.0f, 2.0f, 3.0f, 4.0f, 5.0f, 6.0f, 7.0f, 8.0f, 0.0/0.0, 0.0/0.0 };
+
+    // Specify the number of elements:
+    const int N = 5;
+
+    // Specify the stride length:
+    const int strideX = 2;
+
+    // Initialize a variable for storing the number of non-NaN elements:
+    CBLAS_INT n = 0;
+
+    // Compute the sum:
+    double v = stdlib_strided_dsnannsumors( N, x, strideX, &n );
+
+    // Print the result:
+    printf( "sum: %lf\n", v );
+    printf( "n: %"CBLAS_IFMT"\n", n );
+}
+```
+
+</section>
+
+<!-- /.examples -->
+
+</section>
+
+<!-- /.c -->
+
 <!-- Section for related `stdlib` packages. Do not manually edit this section, as it is automatically populated. -->
 
 <section class="related">

lib/node_modules/@stdlib/blas/ext/base/dsnannsumors/benchmark/benchmark.js

@@ -33,6 +33,19 @@ var dsnannsumors = require( './../lib/dsnannsumors.js' );
 
 // FUNCTIONS //
 
+/**
+* Returns a random number.
+*
+* @private
+* @returns {number} random number
+*/
+function rand() {
+	if ( bernoulli( 0.2 ) ) {
+		return NaN;
+	}
+	return uniform( -10.0, 10.0 );
+}
+
 /**
 * Creates a benchmark function.
 *
@@ -48,13 +61,6 @@ function createBenchmark( len ) {
 	out = new Float64Array( 2 );
 	return benchmark;
 
-	function rand() {
-		if ( bernoulli( 0.2 ) ) {
-			return NaN;
-		}
-		return uniform( -10.0, 10.0 );
-	}
-
 	function benchmark( b ) {
 		var i;
 

lib/node_modules/@stdlib/blas/ext/base/dsnannsumors/benchmark/benchmark.native.js

@@ -42,6 +42,19 @@ var opts = {
 
 // FUNCTIONS //
 
+/**
+* Returns a random number.
+*
+* @private
+* @returns {number} random number
+*/
+function rand() {
+	if ( bernoulli( 0.2 ) ) {
+		return NaN;
+	}
+	return uniform( -10.0, 10.0 );
+}
+
 /**
 * Creates a benchmark function.
 *
@@ -57,13 +70,6 @@ function createBenchmark( len ) {
 	out = new Float64Array( 2 );
 	return benchmark;
 
-	function rand() {
-		if ( bernoulli( 0.2 ) ) {
-			return NaN;
-		}
-		return uniform( -10.0, 10.0 );
-	}
-
 	function benchmark( b ) {
 		var i;
 

lib/node_modules/@stdlib/blas/ext/base/dsnannsumors/benchmark/benchmark.ndarray.js

@@ -33,6 +33,19 @@ var dsnannsumors = require( './../lib/ndarray.js' );
 
 // FUNCTIONS //
 
+/**
+* Returns a random number.
+*
+* @private
+* @returns {number} random number
+*/
+function rand() {
+	if ( bernoulli( 0.2 ) ) {
+		return NaN;
+	}
+	return uniform( -10.0, 10.0 );
+}
+
 /**
 * Creates a benchmark function.
 *
@@ -48,13 +61,6 @@ function createBenchmark( len ) {
 	out = new Float64Array( 2 );
 	return benchmark;
 
-	function rand() {
-		if ( bernoulli( 0.2 ) ) {
-			return NaN;
-		}
-		return uniform( -10.0, 10.0 );
-	}
-
 	function benchmark( b ) {
 		var i;
 

lib/node_modules/@stdlib/blas/ext/base/dsnannsumors/benchmark/benchmark.ndarray.native.js

@@ -42,6 +42,19 @@ var opts = {
 
 // FUNCTIONS //
 
+/**
+* Returns a random number.
+*
+* @private
+* @returns {number} random number
+*/
+function rand() {
+	if ( bernoulli( 0.2 ) ) {
+		return NaN;
+	}
+	return uniform( -10.0, 10.0 );
+}
+
 /**
 * Creates a benchmark function.
 *
@@ -57,13 +70,6 @@ function createBenchmark( len ) {
 	out = new Float64Array( 2 );
 	return benchmark;
 
-	function rand() {
-		if ( bernoulli( 0.2 ) ) {
-			return NaN;
-		}
-		return uniform( -10.0, 10.0 );
-	}
-
 	function benchmark( b ) {
 		var i;
 

lib/node_modules/@stdlib/blas/ext/base/dsnannsumors/benchmark/c/benchmark.length.c

@@ -94,7 +94,7 @@ static float rand_float( void ) {
 * @param len          array length
 * @return elapsed time in seconds
 */
-static double benchmark( int iterations, int len ) {
+static double benchmark1( int iterations, int len ) {
 	double elapsed;
 	float x[ len ];
 	int64_t n;
@@ -126,6 +126,45 @@ static double benchmark( int iterations, int len ) {
 	return elapsed;
 }
 
+/**
+* Runs a benchmark.
+*
+* @param iterations   number of iterations
+* @param len          array length
+* @return elapsed time in seconds
+*/
+static double benchmark2( int iterations, int len ) {
+	double elapsed;
+	float x[ len ];
+	int64_t n;
+	double v;
+	double t;
+	int i;
+
+	for ( i = 0; i < len; i++ ) {
+		if ( rand_float() < 0.2f ) {
+			x[ i ] = 0.0f / 0.0f; // NaN
+		} else {
+			x[ i ] = ( rand_float() * 20000.0f ) - 10000.0f;
+		}
+	}
+	v = 0.0;
+	n = 0;
+	t = tic();
+	for ( i = 0; i < iterations; i++ ) {
+		v = stdlib_strided_dsnannsumors_ndarray( len, x, 1, 0, &n );
+		if ( v != v || n < 0  ) {
+			printf( "should not return NaN\n" );
+			break;
+		}
+	}
+	elapsed = tic() - t;
+	if ( v != v || n < 0 ) {
+		printf( "should not return NaN\n" );
+	}
+	return elapsed;
+}
+
 /**
 * Main execution sequence.
 */
@@ -148,7 +187,18 @@ int main( void ) {
 		for ( j = 0; j < REPEATS; j++ ) {
 			count += 1;
 			printf( "# c::%s:len=%d\n", NAME, len );
-			elapsed = benchmark( iter, len );
+			elapsed = benchmark1( iter, len );
+			print_results( iter, elapsed );
+			printf( "ok %d benchmark finished\n", count );
+		}
+	}
+	for ( i = MIN; i <= MAX; i++ ) {
+		len = pow( 10, i );
+		iter = ITERATIONS / pow( 10, i-1 );
+		for ( j = 0; j < REPEATS; j++ ) {
+			count += 1;
+			printf( "# c::%s:ndarray:len=%d\n", NAME, len );
+			elapsed = benchmark2( iter, len );
 			print_results( iter, elapsed );
 			printf( "ok %d benchmark finished\n", count );
 		}