infomeasure 0.3.0-beta.1

Information theory measures and entropy calculations for Rust
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

// SPDX-FileCopyrightText: 2025-2026 Carlson Büth <code@cbueth.de>
//
// SPDX-License-Identifier: MIT OR Apache-2.0

use infomeasure::estimators::{Entropy, GlobalValue};
use std::fs::File;
use std::io::Write;
use std::time::{Duration, Instant};

// Import test helper functions
use crate::test_helpers::{generate_random_nd_data, measure_execution_time};

/// Measure the performance of the Gaussian kernel entropy calculation
fn measure_gaussian_kernel_performance(
    size: usize,
    dims: usize,
    bandwidth: f64,
    seed: u64,
) -> Duration {
    // Generate random data
    let data = generate_random_nd_data(size, dims, seed);

    // Measure performance using the centralized function
    measure_execution_time(|| match dims {
        1 => {
            let _ =
                Entropy::nd_kernel_with_type::<1>(data.clone(), "gaussian".to_string(), bandwidth)
                    .global_value();
        }
        2 => {
            let _ =
                Entropy::nd_kernel_with_type::<2>(data.clone(), "gaussian".to_string(), bandwidth)
                    .global_value();
        }
        3 => {
            let _ =
                Entropy::nd_kernel_with_type::<3>(data.clone(), "gaussian".to_string(), bandwidth)
                    .global_value();
        }
        4 => {
            let _ =
                Entropy::nd_kernel_with_type::<4>(data.clone(), "gaussian".to_string(), bandwidth)
                    .global_value();
        }
        8 => {
            let _ =
                Entropy::nd_kernel_with_type::<8>(data.clone(), "gaussian".to_string(), bandwidth)
                    .global_value();
        }
        16 => {
            let _ =
                Entropy::nd_kernel_with_type::<16>(data.clone(), "gaussian".to_string(), bandwidth)
                    .global_value();
        }
        32 => {
            let _ =
                Entropy::nd_kernel_with_type::<32>(data.clone(), "gaussian".to_string(), bandwidth)
                    .global_value();
        }
        _ => panic!("Unsupported number of dimensions: {dims}"),
    })
}

/// Measure the performance of the Box kernel entropy calculation
fn measure_box_kernel_performance(size: usize, dims: usize, bandwidth: f64, seed: u64) -> Duration {
    // Generate random data
    let data = generate_random_nd_data(size, dims, seed);

    // Measure performance
    let start = Instant::now();

    match dims {
        1 => Entropy::nd_kernel_with_type::<1>(data.clone(), "box".to_string(), bandwidth)
            .global_value(),
        2 => Entropy::nd_kernel_with_type::<2>(data.clone(), "box".to_string(), bandwidth)
            .global_value(),
        3 => Entropy::nd_kernel_with_type::<3>(data.clone(), "box".to_string(), bandwidth)
            .global_value(),
        4 => Entropy::nd_kernel_with_type::<4>(data.clone(), "box".to_string(), bandwidth)
            .global_value(),
        8 => Entropy::nd_kernel_with_type::<8>(data.clone(), "box".to_string(), bandwidth)
            .global_value(),
        16 => Entropy::nd_kernel_with_type::<16>(data.clone(), "box".to_string(), bandwidth)
            .global_value(),
        32 => Entropy::nd_kernel_with_type::<32>(data.clone(), "box".to_string(), bandwidth)
            .global_value(),
        _ => panic!("Unsupported number of dimensions: {dims}"),
    };

    start.elapsed()
}

#[ignore]
#[test]
fn test_gaussian_kernel_performance() {
    // Define test parameters
    let sizes = [100, 500, 1000, 5000, 10000];
    let dimensions = [1, 2, 4, 8, 16, 32];
    let bandwidth = 0.5;
    let seed = 42;
    let num_runs = 3; // Reduced number of runs to speed up testing

    // Create a file to store the results
    #[cfg(all(feature = "gpu", feature = "fast_exp"))]
    let mut file = File::create("gaussian_gpu_fast_exp_performance.md").unwrap();
    #[cfg(all(feature = "gpu", not(feature = "fast_exp")))]
    let mut file = File::create("gaussian_gpu_performance.md").unwrap();
    #[cfg(all(not(feature = "gpu"), feature = "fast_exp"))]
    let mut file = File::create("gaussian_fast_exp_performance.md").unwrap();
    #[cfg(all(not(feature = "gpu"), not(feature = "fast_exp")))]
    let mut file = File::create("gaussian_baseline_performance.md").unwrap();

    #[cfg(all(feature = "gpu", feature = "fast_exp"))]
    writeln!(file, "# Gaussian Kernel GPU with Fast Exp Performance\n").unwrap();
    #[cfg(all(feature = "gpu", not(feature = "fast_exp")))]
    writeln!(file, "# Gaussian Kernel GPU Performance\n").unwrap();
    #[cfg(all(not(feature = "gpu"), feature = "fast_exp"))]
    writeln!(file, "# Gaussian Kernel Fast Exp Performance\n").unwrap();
    #[cfg(all(not(feature = "gpu"), not(feature = "fast_exp")))]
    writeln!(file, "# Gaussian Kernel Baseline Performance\n").unwrap();

    writeln!(file, "| Data Size | Dimensions | Time (ms) |").unwrap();
    writeln!(file, "|-----------|------------|-----------|").unwrap();

    for &size in &sizes {
        for &dims in &dimensions {
            // Run multiple times and take the average
            let mut total_duration = Duration::new(0, 0);

            for _ in 0..num_runs {
                let duration = measure_gaussian_kernel_performance(size, dims, bandwidth, seed);
                total_duration += duration;
            }

            let avg_duration = total_duration / num_runs as u32;
            let avg_ms = avg_duration.as_millis();

            println!("Gaussian - Size: {size}, Dims: {dims}, Time: {avg_ms} ms");
            writeln!(file, "| {size} | {dims} | {avg_ms} |").unwrap();
        }
    }

    #[cfg(all(feature = "gpu", feature = "fast_exp"))]
    println!(
        "Gaussian kernel GPU with Fast Exp performance results have been saved to gaussian_gpu_fast_exp_performance.md"
    );
    #[cfg(all(feature = "gpu", not(feature = "fast_exp")))]
    println!(
        "Gaussian kernel GPU performance results have been saved to gaussian_gpu_performance.md"
    );
    #[cfg(all(not(feature = "gpu"), feature = "fast_exp"))]
    println!(
        "Gaussian kernel Fast Exp performance results have been saved to gaussian_fast_exp_performance.md"
    );
    #[cfg(all(not(feature = "gpu"), not(feature = "fast_exp")))]
    println!(
        "Gaussian kernel baseline performance results have been saved to gaussian_baseline_performance.md"
    );
}

#[test]
#[ignore]
fn test_box_kernel_performance() {
    // Define test parameters
    let sizes = [100, 500, 1000, 5000, 10000];
    let dimensions = [1, 2, 4, 8, 16, 32];
    let bandwidth = 0.5;
    let seed = 42;
    let num_runs = 3; // Reduced number of runs to speed up testing

    // Create parent directory once before all file operations
    std::fs::create_dir_all("../../internal").unwrap();

    // Create a file to store the results
    #[cfg(all(feature = "gpu", feature = "fast_exp"))]
    let mut file = File::create("../../internal/box_gpu_fast_exp_performance.md").unwrap();
    #[cfg(all(feature = "gpu", not(feature = "fast_exp")))]
    let mut file = File::create("../../internal/box_gpu_performance.md").unwrap();
    #[cfg(all(not(feature = "gpu"), feature = "fast_exp"))]
    let mut file = File::create("../../internal/box_fast_exp_performance.md").unwrap();
    #[cfg(all(not(feature = "gpu"), not(feature = "fast_exp")))]
    let mut file = File::create("../../internal/box_baseline_performance.md").unwrap();

    #[cfg(all(feature = "gpu", feature = "fast_exp"))]
    writeln!(file, "# Box Kernel GPU with Fast Exp Performance\n").unwrap();
    #[cfg(all(feature = "gpu", not(feature = "fast_exp")))]
    writeln!(file, "# Box Kernel GPU Performance\n").unwrap();
    #[cfg(all(not(feature = "gpu"), feature = "fast_exp"))]
    writeln!(file, "# Box Kernel Fast Exp Performance\n").unwrap();
    #[cfg(all(not(feature = "gpu"), not(feature = "fast_exp")))]
    writeln!(file, "# Box Kernel Baseline Performance\n").unwrap();

    writeln!(file, "| Data Size | Dimensions | Time (ms) |").unwrap();
    writeln!(file, "|-----------|------------|-----------|").unwrap();

    for &size in &sizes {
        for &dims in &dimensions {
            // Run multiple times and take the average
            let mut total_duration = Duration::new(0, 0);

            for _ in 0..num_runs {
                let duration = measure_box_kernel_performance(size, dims, bandwidth, seed);
                total_duration += duration;
            }

            let avg_duration = total_duration / num_runs as u32;
            let avg_ms = avg_duration.as_millis();

            println!("Box - Size: {size}, Dims: {dims}, Time: {avg_ms} ms");
            writeln!(file, "| {size} | {dims} | {avg_ms} |").unwrap();
        }
    }

    #[cfg(all(feature = "gpu", feature = "fast_exp"))]
    println!(
        "Box kernel GPU with Fast Exp performance results have been saved to box_gpu_fast_exp_performance.md"
    );
    #[cfg(all(feature = "gpu", not(feature = "fast_exp")))]
    println!("Box kernel GPU performance results have been saved to box_gpu_performance.md");
    #[cfg(all(not(feature = "gpu"), feature = "fast_exp"))]
    println!(
        "Box kernel Fast Exp performance results have been saved to box_fast_exp_performance.md"
    );
    #[cfg(all(not(feature = "gpu"), not(feature = "fast_exp")))]
    println!(
        "Box kernel baseline performance results have been saved to box_baseline_performance.md"
    );
}