blob: 8067dcbc08ea888181eb68ab5d7082c09704803d (
plain) (
blame)
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
|
// Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
package com.yahoo.searchlib.aggregation.hll;
import net.jpountz.xxhash.XXHash32;
import net.jpountz.xxhash.XXHashFactory;
import java.nio.ByteBuffer;
import java.util.ArrayList;
import java.util.List;
import java.util.Random;
import java.util.stream.Collectors;
/**
* This benchmarks performs a series of unique counting tests to analyse the HyperLogLog accuracy.
*/
public class HyperLogLogPrecisionBenchmark {
private static final int MAX_VAL = 256_000;
private static final int MAX_ITERATION = 1000;
private static final XXHash32 hashGenerator = XXHashFactory.safeInstance().hash32();
private static final HyperLogLogEstimator estimator = new HyperLogLogEstimator();
private static final Random random = new Random(424242);
public static void main(String[] args) {
System.out.println("Unique count; Average estimated unique count; Normalized standard error; Standard error; Min; Max");
for (int val = 1; val <= MAX_VAL; val *= 2) {
List<Long> samples = new ArrayList<>();
long sumEstimates = 0;
for (int iteration = 0; iteration < MAX_ITERATION; iteration++) {
long sample = estimateUniqueCount(val);
samples.add(sample);
sumEstimates += sample;
}
double average = sumEstimates / (double) MAX_ITERATION;
long min = samples.stream().min(Long::compare).get();
long max = samples.stream().max(Long::compare).get();
double standardDeviation = getStandardDeviation(samples, average);
System.out.printf("%d; %.2f; %.4f; %.4f; %d; %d\n", val, average, standardDeviation / average, standardDeviation, min, max);
}
}
private static double getStandardDeviation(List<Long> samples, double average) {
double sumSquared = 0;
for (long sample : samples) {
sumSquared += Math.pow(sample - average, 2);
}
return Math.sqrt(sumSquared / samples.size());
}
private static long estimateUniqueCount(int nValues) {
SparseSketch sparse = new SparseSketch();
while (sparse.size() < nValues) {
sparse.aggregate(makeHash(random.nextInt()));
}
if (sparse.size() > HyperLogLog.SPARSE_SKETCH_CONVERSION_THRESHOLD) {
NormalSketch normal = new NormalSketch();
normal.aggregate(sparse.data());
return estimator.estimateCount(normal);
} else {
return estimator.estimateCount(sparse);
}
}
private static int makeHash(int value) {
final int seed = 1333337;
byte[] bytes = ByteBuffer.allocate(4).putInt(value).array();
return hashGenerator.hash(bytes, 0, 4, seed);
}
}
|