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
|
// -----------------------------------------------------------------------------------------------------
// Copyright (c) 2006-2020, Knut Reinert & Freie Universität Berlin
// Copyright (c) 2016-2020, Knut Reinert & MPI für molekulare Genetik
// This file may be used, modified and/or redistributed under the terms of the 3-clause BSD-License
// shipped with this file and also available at: https://github.com/seqan/seqan3/blob/master/LICENSE.md
// -----------------------------------------------------------------------------------------------------
#include <benchmark/benchmark.h>
#include <seqan3/alphabet/nucleotide/dna4.hpp>
#include <seqan3/range/views/minimiser_hash.hpp>
#include <seqan3/test/performance/naive_minimiser_hash.hpp>
#include <seqan3/test/performance/sequence_generator.hpp>
#include <seqan3/test/performance/units.hpp>
#ifdef SEQAN3_HAS_SEQAN2
#include <seqan/index.h>
#include <seqan3/test/performance/seqan2_minimiser.h>
#endif // SEQAN3_HAS_SEQAN2
inline benchmark::Counter bp_per_second(size_t const basepairs)
{
return benchmark::Counter(basepairs,
benchmark::Counter::kIsIterationInvariantRate,
benchmark::Counter::OneK::kIs1000);
}
inline seqan3::shape make_gapped_shape(size_t const k)
{
seqan3::shape shape{};
for (size_t i{0}; i < k - 1; ++i)
shape.push_back((i + 1) % 2);
shape.push_back(1u);
shape.push_back(0u);
return shape;
}
static void arguments(benchmark::internal::Benchmark* b)
{
for (int32_t sequence_length : {50'000, /*1'000'000*/})
{
for (int32_t k : {8, /*16, 24,*/ 30})
{
for (int32_t w : {k + 5, k + 20})
{
b->Args({sequence_length, k, w});
}
}
}
}
enum class method_tag
{
seqan3_ungapped,
seqan3_gapped,
naive,
seqan2_ungapped,
seqan2_gapped
};
#ifdef SEQAN3_HAS_SEQAN2
inline auto make_gapped_shape_seqan2(size_t const k)
{
seqan::String<char> bitmap;
for (size_t i{0}; i < k - 1; ++i)
seqan::append(bitmap, seqan::CharString(std::to_string((i + 1) % 2)));
seqan::append(bitmap, seqan::CharString("1"));
return seqan::Shape<seqan::Dna, seqan::GenericShape>(bitmap);
}
#endif // SEQAN3_HAS_SEQAN2
template <method_tag tag>
void compute_minimisers(benchmark::State & state)
{
auto sequence_length = state.range(0);
size_t k = static_cast<size_t>(state.range(1));
uint32_t w = static_cast<size_t>(state.range(2));
assert(sequence_length > 0);
assert(k > 0);
assert(w > k);
auto seq = seqan3::test::generate_sequence<seqan3::dna4>(sequence_length, 0, 0);
size_t sum{0};
for (auto _ : state)
{
if constexpr (tag == method_tag::naive)
{
for (auto h : seq | seqan3::views::naive_minimiser_hash(seqan3::ungapped{static_cast<uint8_t>(k)}, w))
benchmark::DoNotOptimize(sum += h);
}
else if constexpr (tag == method_tag::seqan3_ungapped)
{
for (auto h : seq | seqan3::views::minimiser_hash(seqan3::ungapped{static_cast<uint8_t>(k)}, seqan3::window_size{w}))
benchmark::DoNotOptimize(sum += h);
}
else if constexpr (tag == method_tag::seqan3_gapped)
{
for (auto h : seq | seqan3::views::minimiser_hash(make_gapped_shape(k), seqan3::window_size{w}))
benchmark::DoNotOptimize(sum += h);
}
#ifdef SEQAN3_HAS_SEQAN2
else
{
auto seqan2_seq = seqan3::test::generate_sequence_seqan2<seqan::Dna>(sequence_length, 0, 0);
using shape_t = std::conditional_t<tag == method_tag::seqan2_ungapped,
seqan::Shape<seqan::Dna, seqan::SimpleShape>,
seqan::Shape<seqan::Dna, seqan::GenericShape>>;
shape_t shape;
if constexpr (tag == method_tag::seqan2_ungapped)
seqan::resize(shape, k);
else
shape = make_gapped_shape_seqan2(k);
minimiser<decltype(shape)> seqan_minimiser(window{w}, kmer{k}, shape);
seqan_minimiser.compute(seqan2_seq);
for (auto h : seqan_minimiser.minimiser_hash)
benchmark::DoNotOptimize(sum += h);
}
#endif // SEQAN3_HAS_SEQAN2
}
state.counters["Throughput[bp/s]"] = bp_per_second(sequence_length - k + 1);
}
template <method_tag tag>
void compute_minimisers_on_poly_A_sequence(benchmark::State & state)
{
auto sequence_length = state.range(0);
size_t k = static_cast<size_t>(state.range(1));
uint32_t w = static_cast<size_t>(state.range(2));
assert(sequence_length > 0);
assert(k > 0);
assert(w > k);
auto seq = std::vector<seqan3::dna4>(sequence_length);
size_t sum{0};
for (auto _ : state)
{
if constexpr (tag == method_tag::seqan3_ungapped)
{
for (auto h : seq | seqan3::views::minimiser_hash(seqan3::ungapped{static_cast<uint8_t>(k)}, seqan3::window_size{w}))
benchmark::DoNotOptimize(sum += h);
}
else if constexpr (tag == method_tag::seqan3_gapped)
{
for (auto h : seq | seqan3::views::minimiser_hash(make_gapped_shape(k), seqan3::window_size{w}))
benchmark::DoNotOptimize(sum += h);
}
}
state.counters["Throughput[bp/s]"] = bp_per_second(sequence_length - k + 1);
}
#ifdef SEQAN3_HAS_SEQAN2
BENCHMARK_TEMPLATE(compute_minimisers, method_tag::seqan2_ungapped)->Apply(arguments);
BENCHMARK_TEMPLATE(compute_minimisers, method_tag::seqan2_gapped)->Apply(arguments);
#endif // SEQAN3_HAS_SEQAN2
BENCHMARK_TEMPLATE(compute_minimisers, method_tag::naive)->Apply(arguments);
BENCHMARK_TEMPLATE(compute_minimisers, method_tag::seqan3_ungapped)->Apply(arguments);
BENCHMARK_TEMPLATE(compute_minimisers, method_tag::seqan3_gapped)->Apply(arguments);
BENCHMARK_TEMPLATE(compute_minimisers_on_poly_A_sequence, method_tag::seqan3_ungapped)->Apply(arguments);
BENCHMARK_TEMPLATE(compute_minimisers_on_poly_A_sequence, method_tag::seqan3_gapped)->Apply(arguments);
BENCHMARK_MAIN();
|
