numrs2 0.3.3

A Rust implementation inspired by NumPy for numerical computing (NumRS2)
Documentation 
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
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330

#!/usr/bin/env python3
"""
NumPy Benchmark Suite for comparison with NumRS2

This script runs various NumPy operations and times them to provide
reference performance data to compare against NumRS2's equivalent operations.

Results are saved to a JSON file that can be loaded by the Rust benchmark suite.
"""

import numpy as np
import json
import os
import time
import sys
from datetime import datetime

# Sample sizes for benchmarks
SMALL_SIZE = 1_000
MEDIUM_SIZE = 10_000
LARGE_SIZE = 100_000
MATRIX_SMALL = (100, 100)
MATRIX_MEDIUM = (1000, 1000)
MATRIX_LARGE = (5000, 5000)

# Number of iterations to run each benchmark
ITERATIONS = 10

# Directory to save benchmark results
BENCH_DIR = os.path.dirname(os.path.abspath(__file__))
RESULTS_FILE = os.path.join(BENCH_DIR, "numpy_benchmark_results.json")

# System information
def get_system_info():
    """Get information about the system running the benchmarks"""
    import platform
    import psutil
    
    try:
        # Get CPU information
        cpu_info = {
            "processor": platform.processor(),
            "python_version": platform.python_version(),
            "numpy_version": np.__version__,
            "platform": platform.platform(),
            "cpu_count": os.cpu_count(),
            "physical_memory_gb": round(psutil.virtual_memory().total / (1024**3), 2)
        }
        return cpu_info
    except ImportError:
        # If psutil is not installed
        return {
            "processor": platform.processor(),
            "python_version": platform.python_version(),
            "numpy_version": np.__version__,
            "platform": platform.platform(),
            "cpu_count": os.cpu_count()
        }

# Benchmark timing function
def time_function(func, *args, **kwargs):
    """
    Time a function execution for multiple iterations and return statistics
    
    Returns:
        dict: Statistics about the timing (mean, min, max, etc.)
    """
    times = []
    
    # Warmup
    func(*args, **kwargs)
    
    # Timed runs
    for _ in range(ITERATIONS):
        start = time.perf_counter()
        func(*args, **kwargs)
        end = time.perf_counter()
        times.append((end - start) * 1000)  # Convert to milliseconds
    
    return {
        "mean": sum(times) / len(times),
        "min": min(times),
        "max": max(times),
        "median": sorted(times)[len(times) // 2],
        "iterations": ITERATIONS
    }

# Benchmark functions
def benchmark_array_creation():
    """Benchmark array creation operations"""
    results = {}
    
    # Time array creation with different sizes
    results["zeros_small"] = time_function(np.zeros, SMALL_SIZE)
    results["zeros_medium"] = time_function(np.zeros, MEDIUM_SIZE)
    results["zeros_large"] = time_function(np.zeros, LARGE_SIZE)
    
    results["ones_small"] = time_function(np.ones, SMALL_SIZE)
    results["ones_medium"] = time_function(np.ones, MEDIUM_SIZE)
    results["ones_large"] = time_function(np.ones, LARGE_SIZE)
    
    results["random_small"] = time_function(np.random.random, (SMALL_SIZE,))
    results["random_medium"] = time_function(np.random.random, (MEDIUM_SIZE,))
    results["random_large"] = time_function(np.random.random, (LARGE_SIZE,))
    
    # 2D arrays
    results["zeros_matrix_small"] = time_function(np.zeros, MATRIX_SMALL)
    results["zeros_matrix_medium"] = time_function(np.zeros, MATRIX_MEDIUM)
    
    results["ones_matrix_small"] = time_function(np.ones, MATRIX_SMALL)
    results["ones_matrix_medium"] = time_function(np.ones, MATRIX_MEDIUM)
    
    results["random_matrix_small"] = time_function(np.random.random, MATRIX_SMALL)
    results["random_matrix_medium"] = time_function(np.random.random, MATRIX_MEDIUM)
    
    # Special array creations
    results["identity_small"] = time_function(np.identity, 100)
    results["identity_medium"] = time_function(np.identity, 1000)
    
    results["arange_small"] = time_function(np.arange, SMALL_SIZE)
    results["arange_medium"] = time_function(np.arange, MEDIUM_SIZE)
    results["arange_large"] = time_function(np.arange, LARGE_SIZE)
    
    return results

def benchmark_array_operations():
    """Benchmark array operations"""
    results = {}
    
    # Create arrays for operations
    a_small = np.random.random(SMALL_SIZE)
    a_medium = np.random.random(MEDIUM_SIZE)
    a_large = np.random.random(LARGE_SIZE)
    
    b_small = np.random.random(SMALL_SIZE)
    b_medium = np.random.random(MEDIUM_SIZE)
    b_large = np.random.random(LARGE_SIZE)
    
    # Basic operations
    results["add_small"] = time_function(np.add, a_small, b_small)
    results["add_medium"] = time_function(np.add, a_medium, b_medium)
    results["add_large"] = time_function(np.add, a_large, b_large)
    
    results["multiply_small"] = time_function(np.multiply, a_small, b_small)
    results["multiply_medium"] = time_function(np.multiply, a_medium, b_medium)
    results["multiply_large"] = time_function(np.multiply, a_large, b_large)
    
    results["sqrt_small"] = time_function(np.sqrt, a_small)
    results["sqrt_medium"] = time_function(np.sqrt, a_medium)
    results["sqrt_large"] = time_function(np.sqrt, a_large)
    
    results["exp_small"] = time_function(np.exp, a_small)
    results["exp_medium"] = time_function(np.exp, a_medium)
    results["exp_large"] = time_function(np.exp, a_large)
    
    results["log_small"] = time_function(np.log, a_small)
    results["log_medium"] = time_function(np.log, a_medium)
    results["log_large"] = time_function(np.log, a_large)
    
    # Reductions
    results["sum_small"] = time_function(np.sum, a_small)
    results["sum_medium"] = time_function(np.sum, a_medium)
    results["sum_large"] = time_function(np.sum, a_large)
    
    results["mean_small"] = time_function(np.mean, a_small)
    results["mean_medium"] = time_function(np.mean, a_medium)
    results["mean_large"] = time_function(np.mean, a_large)
    
    results["std_small"] = time_function(np.std, a_small)
    results["std_medium"] = time_function(np.std, a_medium)
    results["std_large"] = time_function(np.std, a_large)
    
    # Array manipulation
    results["reshape_small"] = time_function(np.reshape, a_small, (SMALL_SIZE // 10, 10))
    results["reshape_medium"] = time_function(np.reshape, a_medium, (MEDIUM_SIZE // 100, 100))
    results["reshape_large"] = time_function(np.reshape, a_large, (LARGE_SIZE // 1000, 1000))
    
    results["transpose_small"] = time_function(lambda x: x.T, np.reshape(a_small, (SMALL_SIZE // 10, 10)))
    results["transpose_medium"] = time_function(lambda x: x.T, np.reshape(a_medium, (MEDIUM_SIZE // 100, 100)))
    
    return results

def benchmark_linear_algebra():
    """Benchmark linear algebra operations"""
    results = {}
    
    # Create matrices for operations
    A_small = np.random.random(MATRIX_SMALL)
    A_medium = np.random.random(MATRIX_MEDIUM)
    
    B_small = np.random.random(MATRIX_SMALL)
    B_medium = np.random.random(MATRIX_MEDIUM)
    
    # Square matrices for decompositions
    S_small = np.random.random((100, 100))
    S_medium = np.random.random((500, 500))
    
    # Matrix operations
    results["matmul_small"] = time_function(np.matmul, A_small, B_small)
    results["matmul_medium"] = time_function(np.matmul, A_medium[:100, :100], B_medium[:100, :100])
    
    results["dot_small"] = time_function(np.dot, A_small.flatten()[:100], B_small.flatten()[:100])
    results["dot_medium"] = time_function(np.dot, A_medium.flatten()[:1000], B_medium.flatten()[:1000])
    
    # Matrix decompositions
    results["svd_small"] = time_function(np.linalg.svd, S_small)
    results["svd_medium"] = time_function(np.linalg.svd, S_medium)
    
    results["qr_small"] = time_function(np.linalg.qr, S_small)
    results["qr_medium"] = time_function(np.linalg.qr, S_medium)
    
    results["cholesky_small"] = time_function(np.linalg.cholesky, S_small @ S_small.T)  # Ensure positive definite
    results["cholesky_medium"] = time_function(np.linalg.cholesky, S_medium @ S_medium.T)
    
    results["det_small"] = time_function(np.linalg.det, S_small)
    results["det_medium"] = time_function(np.linalg.det, S_medium)
    
    results["inv_small"] = time_function(np.linalg.inv, S_small)
    results["inv_medium"] = time_function(np.linalg.inv, S_medium)
    
    return results

def benchmark_distributions():
    """Benchmark random distributions"""
    results = {}
    
    # Basic distributions
    results["normal_small"] = time_function(np.random.normal, 0, 1, SMALL_SIZE)
    results["normal_medium"] = time_function(np.random.normal, 0, 1, MEDIUM_SIZE)
    results["normal_large"] = time_function(np.random.normal, 0, 1, LARGE_SIZE)
    
    results["uniform_small"] = time_function(np.random.uniform, 0, 1, SMALL_SIZE)
    results["uniform_medium"] = time_function(np.random.uniform, 0, 1, MEDIUM_SIZE)
    results["uniform_large"] = time_function(np.random.uniform, 0, 1, LARGE_SIZE)
    
    results["exponential_small"] = time_function(np.random.exponential, 1, SMALL_SIZE)
    results["exponential_medium"] = time_function(np.random.exponential, 1, MEDIUM_SIZE)
    results["exponential_large"] = time_function(np.random.exponential, 1, LARGE_SIZE)
    
    results["beta_small"] = time_function(np.random.beta, 2, 5, SMALL_SIZE)
    results["beta_medium"] = time_function(np.random.beta, 2, 5, MEDIUM_SIZE)
    results["beta_large"] = time_function(np.random.beta, 2, 5, LARGE_SIZE)
    
    results["gamma_small"] = time_function(np.random.gamma, 2, 2, SMALL_SIZE)
    results["gamma_medium"] = time_function(np.random.gamma, 2, 2, MEDIUM_SIZE)
    results["gamma_large"] = time_function(np.random.gamma, 2, 2, LARGE_SIZE)
    
    # Import scipy.stats for advanced distributions
    try:
        import scipy.stats as stats
        
        # Advanced distributions
        results["noncentral_chisquare_small"] = time_function(stats.ncx2.rvs, 2, 1, size=SMALL_SIZE)
        results["noncentral_chisquare_medium"] = time_function(stats.ncx2.rvs, 2, 1, size=MEDIUM_SIZE)
        
        results["noncentral_f_small"] = time_function(stats.ncf.rvs, 2, 5, 1, size=SMALL_SIZE)
        results["noncentral_f_medium"] = time_function(stats.ncf.rvs, 2, 5, 1, size=MEDIUM_SIZE)
        
        results["vonmises_small"] = time_function(stats.vonmises.rvs, 1, size=SMALL_SIZE)
        results["vonmises_medium"] = time_function(stats.vonmises.rvs, 1, size=MEDIUM_SIZE)
        
    except ImportError:
        print("SciPy not available. Skipping advanced distribution benchmarks.")
    
    return results

def run_benchmarks():
    """Run all benchmarks and save results to a JSON file"""
    results = {}
    
    print("Running NumPy Benchmarks for comparison with NumRS2")
    print("-" * 50)
    
    # Get system information
    results["system_info"] = get_system_info()
    results["timestamp"] = datetime.now().isoformat()
    
    # Run benchmarks
    print("Benchmarking array creation...")
    results["array_creation"] = benchmark_array_creation()
    
    print("Benchmarking array operations...")
    results["array_operations"] = benchmark_array_operations()
    
    print("Benchmarking linear algebra...")
    results["linear_algebra"] = benchmark_linear_algebra()
    
    print("Benchmarking distributions...")
    results["distributions"] = benchmark_distributions()
    
    # Save results to JSON file
    with open(RESULTS_FILE, 'w') as f:
        json.dump(results, f, indent=2)
    
    print(f"\nBenchmark results saved to {RESULTS_FILE}")
    return results

if __name__ == "__main__":
    run_benchmarks()
    
    # Print summary of results
    with open(RESULTS_FILE, 'r') as f:
        results = json.load(f)
    
    print("\nBenchmark Summary (mean times in ms):")
    print("-" * 50)
    
    # Print a few key results
    print("Array Creation:")
    print(f"  zeros_medium: {results['array_creation']['zeros_medium']['mean']:.2f} ms")
    print(f"  random_medium: {results['array_creation']['random_medium']['mean']:.2f} ms")
    
    print("\nArray Operations:")
    print(f"  add_medium: {results['array_operations']['add_medium']['mean']:.2f} ms")
    print(f"  multiply_medium: {results['array_operations']['multiply_medium']['mean']:.2f} ms")
    print(f"  sqrt_medium: {results['array_operations']['sqrt_medium']['mean']:.2f} ms")
    
    print("\nLinear Algebra:")
    print(f"  matmul_small: {results['linear_algebra']['matmul_small']['mean']:.2f} ms")
    print(f"  svd_small: {results['linear_algebra']['svd_small']['mean']:.2f} ms")
    print(f"  qr_small: {results['linear_algebra']['qr_small']['mean']:.2f} ms")
    
    print("\nDistributions:")
    print(f"  normal_medium: {results['distributions']['normal_medium']['mean']:.2f} ms")
    print(f"  beta_medium: {results['distributions']['beta_medium']['mean']:.2f} ms")
    
    if "noncentral_chisquare_medium" in results["distributions"]:
        print(f"  noncentral_chisquare_medium: {results['distributions']['noncentral_chisquare_medium']['mean']:.2f} ms")
        
    print("\nRun this benchmark on the same system as NumRS2 benchmarks for valid comparisons.")