benchmark

A library to support the benchmarking of functions, similar to unit-tests.

Discussion group: https://groups.google.com/d/forum/benchmark-discuss

IRC channel: https://freenode.net #googlebenchmark

Known issues and common problems

Additional Tooling Documentation

Example usage

Basic usage

Define a function that executes the code to be measured.

static void BM_StringCreation(benchmark::State& state) {
  while (state.KeepRunning())
    std::string empty_string;
}
// Register the function as a benchmark
BENCHMARK(BM_StringCreation);

// Define another benchmark
static void BM_StringCopy(benchmark::State& state) {
  std::string x = "hello";
  while (state.KeepRunning())
    std::string copy(x);
}
BENCHMARK(BM_StringCopy);

BENCHMARK_MAIN();

Passing arguments

Sometimes a family of benchmarks can be implemented with just one routine that takes an extra argument to specify which one of the family of benchmarks to run. For example, the following code defines a family of benchmarks for measuring the speed of memcpy() calls of different lengths:

static void BM_memcpy(benchmark::State& state) {
  char* src = new char[state.range(0)];
  char* dst = new char[state.range(0)];
  memset(src, 'x', state.range(0));
  while (state.KeepRunning())
    memcpy(dst, src, state.range(0));
  state.SetBytesProcessed(int64_t(state.iterations()) *
                          int64_t(state.range(0)));
  delete[] src;
  delete[] dst;
}
BENCHMARK(BM_memcpy)->Arg(8)->Arg(64)->Arg(512)->Arg(1<<10)->Arg(8<<10);

The preceding code is quite repetitive, and can be replaced with the following short-hand. The following invocation will pick a few appropriate arguments in the specified range and will generate a benchmark for each such argument.

BENCHMARK(BM_memcpy)->Range(8, 8<<10);

By default the arguments in the range are generated in multiples of eight and the command above selects [ 8, 64, 512, 4k, 8k ]. In the following code the range multiplier is changed to multiples of two.

BENCHMARK(BM_memcpy)->RangeMultiplier(2)->Range(8, 8<<10);

Now arguments generated are [ 8, 16, 32, 64, 128, 256, 512, 1024, 2k, 4k, 8k ].

You might have a benchmark that depends on two or more inputs. For example, the following code defines a family of benchmarks for measuring the speed of set insertion.

static void BM_SetInsert(benchmark::State& state) {
  while (state.KeepRunning()) {
    state.PauseTiming();
    std::set<int> data = ConstructRandomSet(state.range(0));
    state.ResumeTiming();
    for (int j = 0; j < state.range(1); ++j)
      data.insert(RandomNumber());
  }
}
BENCHMARK(BM_SetInsert)
    ->Args({1<<10, 1})
    ->Args({1<<10, 8})
    ->Args({1<<10, 64})
    ->Args({1<<10, 512})
    ->Args({8<<10, 1})
    ->Args({8<<10, 8})
    ->Args({8<<10, 64})
    ->Args({8<<10, 512});

The preceding code is quite repetitive, and can be replaced with the following short-hand. The following macro will pick a few appropriate arguments in the product of the two specified ranges and will generate a benchmark for each such pair.

BENCHMARK(BM_SetInsert)->Ranges({{1<<10, 8<<10}, {1, 512}});

For more complex patterns of inputs, passing a custom function to Apply allows programmatic specification of an arbitrary set of arguments on which to run the benchmark. The following example enumerates a dense range on one parameter, and a sparse range on the second.

static void CustomArguments(benchmark::internal::Benchmark* b) {
  for (int i = 0; i <= 10; ++i)
    for (int j = 32; j <= 1024*1024; j *= 8)
      b->Args({i, j});
}
BENCHMARK(BM_SetInsert)->Apply(CustomArguments);

Calculate asymptotic complexity (Big O)

Asymptotic complexity might be calculated for a family of benchmarks. The following code will calculate the coefficient for the high-order term in the running time and the normalized root-mean square error of string comparison.

static void BM_StringCompare(benchmark::State& state) {
  std::string s1(state.range(0), '-');
  std::string s2(state.range(0), '-');
  while (state.KeepRunning()) {
    benchmark::DoNotOptimize(s1.compare(s2));
  }
  state.SetComplexityN(state.range(0));
}
BENCHMARK(BM_StringCompare)
    ->RangeMultiplier(2)->Range(1<<10, 1<<18)->Complexity(benchmark::oN);

As shown in the following invocation, asymptotic complexity might also be calculated automatically.

BENCHMARK(BM_StringCompare)
    ->RangeMultiplier(2)->Range(1<<10, 1<<18)->Complexity();

The following code will specify asymptotic complexity with a lambda function, that might be used to customize high-order term calculation.

BENCHMARK(BM_StringCompare)->RangeMultiplier(2)
    ->Range(1<<10, 1<<18)->Complexity([](int n)->double{return n; });

Templated benchmarks

Templated benchmarks work the same way: This example produces and consumes messages of size sizeof(v) range_x times. It also outputs throughput in the absence of multiprogramming.

template <class Q> int BM_Sequential(benchmark::State& state) {
  Q q;
  typename Q::value_type v;
  while (state.KeepRunning()) {
    for (int i = state.range(0); i--; )
      q.push(v);
    for (int e = state.range(0); e--; )
      q.Wait(&v);
  }
  // actually messages, not bytes:
  state.SetBytesProcessed(
      static_cast<int64_t>(state.iterations())*state.range(0));
}
BENCHMARK_TEMPLATE(BM_Sequential, WaitQueue<int>)->Range(1<<0, 1<<10);

Three macros are provided for adding benchmark templates.

#if __cplusplus >= 201103L // C++11 and greater.
#define BENCHMARK_TEMPLATE(func, ...) // Takes any number of parameters.
#else // C++ < C++11
#define BENCHMARK_TEMPLATE(func, arg1)
#endif
#define BENCHMARK_TEMPLATE1(func, arg1)
#define BENCHMARK_TEMPLATE2(func, arg1, arg2)

Passing arbitrary arguments to a benchmark

In C++11 it is possible to define a benchmark that takes an arbitrary number of extra arguments. The BENCHMARK_CAPTURE(func, test_case_name, ...args) macro creates a benchmark that invokes func with the benchmark::State as the first argument followed by the specified args.... The test_case_name is appended to the name of the benchmark and should describe the values passed.

template <class ...ExtraArgs>`
void BM_takes_args(benchmark::State& state, ExtraArgs&&... extra_args) {
  [...]
}
// Registers a benchmark named "BM_takes_args/int_string_test` that passes
// the specified values to `extra_args`.
BENCHMARK_CAPTURE(BM_takes_args, int_string_test, 42, std::string("abc"));

Note that elements of ...args may refer to global variables. Users should avoid modifying global state inside of a benchmark.

Using RegisterBenchmark(name, fn, args...)

The RegisterBenchmark(name, func, args...) function provides an alternative way to create and register benchmarks. RegisterBenchmark(name, func, args...) creates, registers, and returns a pointer to a new benchmark with the specified name that invokes func(st, args...) where st is a benchmark::State object.

Unlike the BENCHMARK registration macros, which can only be used at the global scope, the RegisterBenchmark can be called anywhere. This allows for benchmark tests to be registered programmatically.

Additionally RegisterBenchmark allows any callable object to be registered as a benchmark. Including capturing lambdas and function objects. This allows the creation

For Example:

auto BM_test = [](benchmark::State& st, auto Inputs) { /* ... */ };

int main(int argc, char** argv) {
  for (auto& test_input : { /* ... */ })
      benchmark::RegisterBenchmark(test_input.name(), BM_test, test_input);
  benchmark::Initialize(&argc, argv);
  benchmark::RunSpecifiedBenchmarks();
}

Multithreaded benchmarks

In a multithreaded test (benchmark invoked by multiple threads simultaneously), it is guaranteed that none of the threads will start until all have called KeepRunning, and all will have finished before KeepRunning returns false. As such, any global setup or teardown can be wrapped in a check against the thread index:

static void BM_MultiThreaded(benchmark::State& state) {
  if (state.thread_index == 0) {
    // Setup code here.
  }
  while (state.KeepRunning()) {
    // Run the test as normal.
  }
  if (state.thread_index == 0) {
    // Teardown code here.
  }
}
BENCHMARK(BM_MultiThreaded)->Threads(2);

If the benchmarked code itself uses threads and you want to compare it to single-threaded code, you may want to use real-time ("wallclock") measurements for latency comparisons:

BENCHMARK(BM_test)->Range(8, 8<<10)->UseRealTime();

Without UseRealTime, CPU time is used by default.

Manual timing

For benchmarking something for which neither CPU time nor real-time are correct or accurate enough, completely manual timing is supported using the UseManualTime function.

When UseManualTime is used, the benchmarked code must call SetIterationTime once per iteration of the KeepRunning loop to report the manually measured time.

An example use case for this is benchmarking GPU execution (e.g. OpenCL or CUDA kernels, OpenGL or Vulkan or Direct3D draw calls), which cannot be accurately measured using CPU time or real-time. Instead, they can be measured accurately using a dedicated API, and these measurement results can be reported back with SetIterationTime.

static void BM_ManualTiming(benchmark::State& state) {
  int microseconds = state.range(0);
  std::chrono::duration<double, std::micro> sleep_duration {
    static_cast<double>(microseconds)
  };

  while (state.KeepRunning()) {
    auto start = std::chrono::high_resolution_clock::now();
    // Simulate some useful workload with a sleep
    std::this_thread::sleep_for(sleep_duration);
    auto end   = std::chrono::high_resolution_clock::now();

    auto elapsed_seconds =
      std::chrono::duration_cast<std::chrono::duration<double>>(
        end - start);

    state.SetIterationTime(elapsed_seconds.count());
  }
}
BENCHMARK(BM_ManualTiming)->Range(1, 1<<17)->UseManualTime();

Preventing optimisation

To prevent a value or expression from being optimized away by the compiler the benchmark::DoNotOptimize(...) and benchmark::ClobberMemory() functions can be used.

static void BM_test(benchmark::State& state) {
  while (state.KeepRunning()) {
      int x = 0;
      for (int i=0; i < 64; ++i) {
        benchmark::DoNotOptimize(x += i);
      }
  }
}

DoNotOptimize(<expr>) forces the result of <expr> to be stored in either memory or a register. For GNU based compilers it acts as read/write barrier for global memory. More specifically it forces the compiler to flush pending writes to memory and reload any other values as necessary.

Note that DoNotOptimize(<expr>) does not prevent optimizations on <expr> in any way. <expr> may even be removed entirely when the result is already known. For example:

  /* Example 1: `<expr>` is removed entirely. */
  int foo(int x) { return x + 42; }
  while (...) DoNotOptimize(foo(0)); // Optimized to DoNotOptimize(42);

  /*  Example 2: Result of '<expr>' is only reused */
  int bar(int) __attribute__((const));
  while (...) DoNotOptimize(bar(0)); // Optimized to:
  // int __result__ = bar(0);
  // while (...) DoNotOptimize(__result__);

The second tool for preventing optimizations is ClobberMemory(). In essence ClobberMemory() forces the compiler to perform all pending writes to global memory. Memory managed by block scope objects must be "escaped" using DoNotOptimize(...) before it can be clobbered. In the below example ClobberMemory() prevents the call to v.push_back(42) from being optimized away.

static void BM_vector_push_back(benchmark::State& state) {
  while (state.KeepRunning()) {
    std::vector<int> v;
    v.reserve(1);
    benchmark::DoNotOptimize(v.data()); // Allow v.data() to be clobbered.
    v.push_back(42);
    benchmark::ClobberMemory(); // Force 42 to be written to memory.
  }
}

Note that ClobberMemory() is only available for GNU or MSVC based compilers.

Set time unit manually

If a benchmark runs a few milliseconds it may be hard to visually compare the measured times, since the output data is given in nanoseconds per default. In order to manually set the time unit, you can specify it manually:

BENCHMARK(BM_test)->Unit(benchmark::kMillisecond);

Controlling number of iterations

In all cases, the number of iterations for which the benchmark is run is governed by the amount of time the benchmark takes. Concretely, the number of iterations is at least one, not more than 1e9, until CPU time is greater than the minimum time, or the wallclock time is 5x minimum time. The minimum time is set as a flag --benchmark_min_time or per benchmark by calling MinTime on the registered benchmark object.

Reporting the mean and standard devation by repeated benchmarks

By default each benchmark is run once and that single result is reported. However benchmarks are often noisy and a single result may not be representative of the overall behavior. For this reason it's possible to repeatedly rerun the benchmark.

The number of runs of each benchmark is specified globally by the --benchmark_repetitions flag or on a per benchmark basis by calling Repetitions on the registered benchmark object. When a benchmark is run more than once the mean and standard deviation of the runs will be reported.

Additionally the --benchmark_report_aggregates_only={true|false} flag or ReportAggregatesOnly(bool) function can be used to change how repeated tests are reported. By default the result of each repeated run is reported. When this option is 'true' only the mean and standard deviation of the runs is reported. Calling ReportAggregatesOnly(bool) on a registered benchmark object overrides the value of the flag for that benchmark.

Fixtures

Fixture tests are created by first defining a type that derives from ::benchmark::Fixture and then creating/registering the tests using the following macros:

• BENCHMARK_F(ClassName, Method)
• BENCHMARK_DEFINE_F(ClassName, Method)
• BENCHMARK_REGISTER_F(ClassName, Method)

For Example:

class MyFixture : public benchmark::Fixture {};

BENCHMARK_F(MyFixture, FooTest)(benchmark::State& st) {
   while (st.KeepRunning()) {
     ...
  }
}

BENCHMARK_DEFINE_F(MyFixture, BarTest)(benchmark::State& st) {
   while (st.KeepRunning()) {
     ...
  }
}
/* BarTest is NOT registered */
BENCHMARK_REGISTER_F(MyFixture, BarTest)->Threads(2);
/* BarTest is now registered */

User-defined counters

You can add your own counters with user-defined names. The example below will add columns "Foo", "Bar" and "Baz" in its output:

static void UserCountersExample1(benchmark::State& state) {
  double numFoos = 0, numBars = 0, numBazs = 0;
  while (state.KeepRunning()) {
    // ... count Foo,Bar,Baz events
  }
  state.counters["Foo"] = numFoos;
  state.counters["Bar"] = numBars;
  state.counters["Baz"] = numBazs;
}

The state.counters object is a std::map with std::string keys and Counter values. The latter is a double-like class, via an implicit conversion to double&. Thus you can use all of the standard arithmetic assignment operators (=,+=,-=,*=,/=) to change the value of each counter.

In multithreaded benchmarks, each counter is set on the calling thread only. When the benchmark finishes, the counters from each thread will be summed; the resulting sum is the value which will be shown for the benchmark.

The Counter constructor accepts two parameters: the value as a double and a bit flag which allows you to show counters as rates and/or as per-thread averages:

  // sets a simple counter
  state.counters["Foo"] = numFoos;

  // Set the counter as a rate. It will be presented divided
  // by the duration of the benchmark.
  state.counters["FooRate"] = Counter(numFoos, benchmark::Counter::kIsRate);

  // Set the counter as a thread-average quantity. It will
  // be presented divided by the number of threads.
  state.counters["FooAvg"] = Counter(numFoos, benchmark::Counter::kAvgThreads);

  // There's also a combined flag:
  state.counters["FooAvgRate"] = Counter(numFoos,benchmark::Counter::kAvgThreadsRate);

When you're compiling in C++11 mode or later you can use insert() with std::initializer_list:

  // With C++11, this can be done:
  state.counters.insert({{"Foo", numFoos}, {"Bar", numBars}, {"Baz", numBazs}});
  // ... instead of:
  state.counters["Foo"] = numFoos;
  state.counters["Bar"] = numBars;
  state.counters["Baz"] = numBazs;

Counter reporting

When using the console reporter, by default, user counters are are printed at the end after the table, the same way as bytes_processed and items_processed. This is best for cases in which there are few counters, or where there are only a couple of lines per benchmark. Here's an example of the default output:

------------------------------------------------------------------------------
Benchmark                        Time           CPU Iterations UserCounters...
------------------------------------------------------------------------------
BM_UserCounter/threads:8      2248 ns      10277 ns      68808 Bar=16 Bat=40 Baz=24 Foo=8
BM_UserCounter/threads:1      9797 ns       9788 ns      71523 Bar=2 Bat=5 Baz=3 Foo=1024m
BM_UserCounter/threads:2      4924 ns       9842 ns      71036 Bar=4 Bat=10 Baz=6 Foo=2
BM_UserCounter/threads:4      2589 ns      10284 ns      68012 Bar=8 Bat=20 Baz=12 Foo=4
BM_UserCounter/threads:8      2212 ns      10287 ns      68040 Bar=16 Bat=40 Baz=24 Foo=8
BM_UserCounter/threads:16     1782 ns      10278 ns      68144 Bar=32 Bat=80 Baz=48 Foo=16
BM_UserCounter/threads:32     1291 ns      10296 ns      68256 Bar=64 Bat=160 Baz=96 Foo=32
BM_UserCounter/threads:4      2615 ns      10307 ns      68040 Bar=8 Bat=20 Baz=12 Foo=4
BM_Factorial                    26 ns         26 ns   26608979 40320
BM_Factorial/real_time          26 ns         26 ns   26587936 40320
BM_CalculatePiRange/1           16 ns         16 ns   45704255 0
BM_CalculatePiRange/8           73 ns         73 ns    9520927 3.28374
BM_CalculatePiRange/64         609 ns        609 ns    1140647 3.15746
BM_CalculatePiRange/512       4900 ns       4901 ns     142696 3.14355

If this doesn't suit you, you can print each counter as a table column by passing the flag --benchmark_counters_tabular=true to the benchmark application. This is best for cases in which there are a lot of counters, or a lot of lines per individual benchmark. Note that this will trigger a reprinting of the table header any time the counter set changes between individual benchmarks. Here's an example of corresponding output when --benchmark_counters_tabular=true is passed:

---------------------------------------------------------------------------------------
Benchmark                        Time           CPU Iterations    Bar   Bat   Baz   Foo
---------------------------------------------------------------------------------------
BM_UserCounter/threads:8      2198 ns       9953 ns      70688     16    40    24     8
BM_UserCounter/threads:1      9504 ns       9504 ns      73787      2     5     3     1
BM_UserCounter/threads:2      4775 ns       9550 ns      72606      4    10     6     2
BM_UserCounter/threads:4      2508 ns       9951 ns      70332      8    20    12     4
BM_UserCounter/threads:8      2055 ns       9933 ns      70344     16    40    24     8
BM_UserCounter/threads:16     1610 ns       9946 ns      70720     32    80    48    16
BM_UserCounter/threads:32     1192 ns       9948 ns      70496     64   160    96    32
BM_UserCounter/threads:4      2506 ns       9949 ns      70332      8    20    12     4
--------------------------------------------------------------
Benchmark                        Time           CPU Iterations
--------------------------------------------------------------
BM_Factorial                    26 ns         26 ns   26392245 40320
BM_Factorial/real_time          26 ns         26 ns   26494107 40320
BM_CalculatePiRange/1           15 ns         15 ns   45571597 0
BM_CalculatePiRange/8           74 ns         74 ns    9450212 3.28374
BM_CalculatePiRange/64         595 ns        595 ns    1173901 3.15746
BM_CalculatePiRange/512       4752 ns       4752 ns     147380 3.14355
BM_CalculatePiRange/4k       37970 ns      37972 ns      18453 3.14184
BM_CalculatePiRange/32k     303733 ns     303744 ns       2305 3.14162
BM_CalculatePiRange/256k   2434095 ns    2434186 ns        288 3.1416
BM_CalculatePiRange/1024k  9721140 ns    9721413 ns         71 3.14159
BM_CalculatePi/threads:8      2255 ns       9943 ns      70936

Note above the additional header printed when the benchmark changes from BM_UserCounter to BM_Factorial. This is because BM_Factorial does not have the same counter set as BM_UserCounter.

Exiting Benchmarks in Error

When errors caused by external influences, such as file I/O and network communication, occur within a benchmark the State::SkipWithError(const char* msg) function can be used to skip that run of benchmark and report the error. Note that only future iterations of the KeepRunning() are skipped. Users may explicitly return to exit the benchmark immediately.

The SkipWithError(...) function may be used at any point within the benchmark, including before and after the KeepRunning() loop.

For example:

static void BM_test(benchmark::State& state) {
  auto resource = GetResource();
  if (!resource.good()) {
      state.SkipWithError("Resource is not good!");
      // KeepRunning() loop will not be entered.
  }
  while (state.KeepRunning()) {
      auto data = resource.read_data();
      if (!resource.good()) {
        state.SkipWithError("Failed to read data!");
        break; // Needed to skip the rest of the iteration.
     }
     do_stuff(data);
  }
}

Running a subset of the benchmarks

The --benchmark_filter=<regex> option can be used to only run the benchmarks which match the specified <regex>. For example:

$ ./run_benchmarks.x --benchmark_filter=BM_memcpy/32
Run on (1 X 2300 MHz CPU )
2016-06-25 19:34:24
Benchmark              Time           CPU Iterations
----------------------------------------------------
BM_memcpy/32          11 ns         11 ns   79545455
BM_memcpy/32k       2181 ns       2185 ns     324074
BM_memcpy/32          12 ns         12 ns   54687500
BM_memcpy/32k       1834 ns       1837 ns     357143

Output Formats

The library supports multiple output formats. Use the --benchmark_format=<console|json|csv> flag to set the format type. console is the default format.

The Console format is intended to be a human readable format. By default the format generates color output. Context is output on stderr and the tabular data on stdout. Example tabular output looks like:

Benchmark                               Time(ns)    CPU(ns) Iterations
----------------------------------------------------------------------
BM_SetInsert/1024/1                        28928      29349      23853  133.097kB/s   33.2742k items/s
BM_SetInsert/1024/8                        32065      32913      21375  949.487kB/s   237.372k items/s
BM_SetInsert/1024/10                       33157      33648      21431  1.13369MB/s   290.225k items/s

The JSON format outputs human readable json split into two top level attributes. The context attribute contains information about the run in general, including information about the CPU and the date. The benchmarks attribute contains a list of ever benchmark run. Example json output looks like:

{
  "context": {
    "date": "2015/03/17-18:40:25",
    "num_cpus": 40,
    "mhz_per_cpu": 2801,
    "cpu_scaling_enabled": false,
    "build_type": "debug"
  },
  "benchmarks": [
    {
      "name": "BM_SetInsert/1024/1",
      "iterations": 94877,
      "real_time": 29275,
      "cpu_time": 29836,
      "bytes_per_second": 134066,
      "items_per_second": 33516
    },
    {
      "name": "BM_SetInsert/1024/8",
      "iterations": 21609,
      "real_time": 32317,
      "cpu_time": 32429,
      "bytes_per_second": 986770,
      "items_per_second": 246693
    },
    {
      "name": "BM_SetInsert/1024/10",
      "iterations": 21393,
      "real_time": 32724,
      "cpu_time": 33355,
      "bytes_per_second": 1199226,
      "items_per_second": 299807
    }
  ]
}

The CSV format outputs comma-separated values. The context is output on stderr and the CSV itself on stdout. Example CSV output looks like:

name,iterations,real_time,cpu_time,bytes_per_second,items_per_second,label
"BM_SetInsert/1024/1",65465,17890.7,8407.45,475768,118942,
"BM_SetInsert/1024/8",116606,18810.1,9766.64,3.27646e+06,819115,
"BM_SetInsert/1024/10",106365,17238.4,8421.53,4.74973e+06,1.18743e+06,

Output Files

The library supports writing the output of the benchmark to a file specified by --benchmark_out=<filename>. The format of the output can be specified using --benchmark_out_format={json|console|csv}. Specifying --benchmark_out does not suppress the console output.

Debug vs Release

By default, benchmark builds as a debug library. You will see a warning in the output when this is the case. To build it as a release library instead, use:

cmake -DCMAKE_BUILD_TYPE=Release

To enable link-time optimisation, use

cmake -DCMAKE_BUILD_TYPE=Release -DBENCHMARK_ENABLE_LTO=true

Linking against the library

When using gcc, it is necessary to link against pthread to avoid runtime exceptions. This is due to how gcc implements std::thread. See issue #67 for more details.

Compiler Support

Google Benchmark uses C++11 when building the library. As such we require a modern C++ toolchain, both compiler and standard library.

The following minimum versions are strongly recommended build the library:

• GCC 4.8
• Clang 3.4
• Visual Studio 2013
• Intel 2015 Update 1

Anything older may work.

Note: Using the library and its headers in C++03 is supported. C++11 is only required to build the library.

Known Issues

Windows

• Users must manually link shlwapi.lib. Failure to do so may result in unresolved symbols.