aboutsummaryrefslogtreecommitdiff
path: root/src/bench/bench.cpp
blob: ac14444eef17f5a7aea21462c153a8314900c1ef (plain)
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
// Copyright (c) 2015-2016 The Bitcoin Core developers
// Distributed under the MIT software license, see the accompanying
// file COPYING or http://www.opensource.org/licenses/mit-license.php.

#include "bench.h"
#include "perf.h"

#include <assert.h>
#include <iostream>
#include <iomanip>
#include <sys/time.h>

benchmark::BenchRunner::BenchmarkMap &benchmark::BenchRunner::benchmarks() {
    static std::map<std::string, benchmark::BenchFunction> benchmarks_map;
    return benchmarks_map;
}

static double gettimedouble(void) {
    struct timeval tv;
    gettimeofday(&tv, nullptr);
    return tv.tv_usec * 0.000001 + tv.tv_sec;
}

benchmark::BenchRunner::BenchRunner(std::string name, benchmark::BenchFunction func)
{
    benchmarks().insert(std::make_pair(name, func));
}

void
benchmark::BenchRunner::RunAll(double elapsedTimeForOne)
{
    perf_init();
    std::cout << "#Benchmark" << "," << "count" << "," << "min" << "," << "max" << "," << "average" << ","
              << "min_cycles" << "," << "max_cycles" << "," << "average_cycles" << "\n";

    for (const auto &p: benchmarks()) {
        State state(p.first, elapsedTimeForOne);
        p.second(state);
    }
    perf_fini();
}

bool benchmark::State::KeepRunning()
{
    if (count & countMask) {
      ++count;
      return true;
    }
    double now;
    uint64_t nowCycles;
    if (count == 0) {
        lastTime = beginTime = now = gettimedouble();
        lastCycles = beginCycles = nowCycles = perf_cpucycles();
    }
    else {
        now = gettimedouble();
        double elapsed = now - lastTime;
        double elapsedOne = elapsed * countMaskInv;
        if (elapsedOne < minTime) minTime = elapsedOne;
        if (elapsedOne > maxTime) maxTime = elapsedOne;

        // We only use relative values, so don't have to handle 64-bit wrap-around specially
        nowCycles = perf_cpucycles();
        uint64_t elapsedOneCycles = (nowCycles - lastCycles) * countMaskInv;
        if (elapsedOneCycles < minCycles) minCycles = elapsedOneCycles;
        if (elapsedOneCycles > maxCycles) maxCycles = elapsedOneCycles;

        if (elapsed*128 < maxElapsed) {
          // If the execution was much too fast (1/128th of maxElapsed), increase the count mask by 8x and restart timing.
          // The restart avoids including the overhead of this code in the measurement.
          countMask = ((countMask<<3)|7) & ((1LL<<60)-1);
          countMaskInv = 1./(countMask+1);
          count = 0;
          minTime = std::numeric_limits<double>::max();
          maxTime = std::numeric_limits<double>::min();
          minCycles = std::numeric_limits<uint64_t>::max();
          maxCycles = std::numeric_limits<uint64_t>::min();
          return true;
        }
        if (elapsed*16 < maxElapsed) {
          uint64_t newCountMask = ((countMask<<1)|1) & ((1LL<<60)-1);
          if ((count & newCountMask)==0) {
              countMask = newCountMask;
              countMaskInv = 1./(countMask+1);
          }
        }
    }
    lastTime = now;
    lastCycles = nowCycles;
    ++count;

    if (now - beginTime < maxElapsed) return true; // Keep going

    --count;

    assert(count != 0 && "count == 0 => (now == 0 && beginTime == 0) => return above");

    // Output results
    double average = (now-beginTime)/count;
    int64_t averageCycles = (nowCycles-beginCycles)/count;
    std::cout << std::fixed << std::setprecision(15) << name << "," << count << "," << minTime << "," << maxTime << "," << average << ","
              << minCycles << "," << maxCycles << "," << averageCycles << "\n";

    return false;
}