Reduce compilation costs of benchmarks

We replaced some simple std::algorithm usage by loops, and reduced
header inclusion.

6 files changed, 53 insertions, 40 deletions

diff --git a/src/catch2/benchmark/catch_execution_plan.hpp b/src/catch2/benchmark/catch_execution_plan.hpp
index 039de7ee..4f60a646 100644
--- a/src/catch2/benchmark/catch_execution_plan.hpp
+++ b/src/catch2/benchmark/catch_execution_plan.hpp
@@ -17,8 +17,7 @@
 #include <catch2/benchmark/detail/catch_repeat.hpp>
 #include <catch2/benchmark/detail/catch_run_for_at_least.hpp>
 
-#include <algorithm>
-#include <iterator>
+#include <vector>
 
 namespace Catch {
     namespace Benchmark {
@@ -41,14 +40,17 @@ namespace Catch {
                 Detail::run_for_at_least<Clock>(std::chrono::duration_cast<ClockDuration<Clock>>(warmup_time), warmup_iterations, Detail::repeat(now<Clock>{}));
 
                 std::vector<FloatDuration<Clock>> times;
-                times.reserve(cfg.benchmarkSamples());
-                std::generate_n(std::back_inserter(times), cfg.benchmarkSamples(), [this, env] {
+                const auto num_samples = cfg.benchmarkSamples();
+                times.reserve( num_samples );
+                for ( size_t i = 0; i < num_samples; ++i ) {
                     Detail::ChronometerModel<Clock> model;
-                    this->benchmark(Chronometer(model, iterations_per_sample));
+                    this->benchmark( Chronometer( model, iterations_per_sample ) );
                     auto sample_time = model.elapsed() - env.clock_cost.mean;
-                    if (sample_time < FloatDuration<Clock>::zero()) sample_time = FloatDuration<Clock>::zero();
-                    return sample_time / iterations_per_sample;
-                });
+                    if ( sample_time < FloatDuration<Clock>::zero() ) {
+                        sample_time = FloatDuration<Clock>::zero();
+                    }
+                    times.push_back(sample_time / iterations_per_sample);
+                }
                 return times;
             }
         };
diff --git a/src/catch2/benchmark/catch_sample_analysis.hpp b/src/catch2/benchmark/catch_sample_analysis.hpp
index d849d246..97b8fe50 100644
--- a/src/catch2/benchmark/catch_sample_analysis.hpp
+++ b/src/catch2/benchmark/catch_sample_analysis.hpp
@@ -10,14 +10,11 @@
 #ifndef CATCH_SAMPLE_ANALYSIS_HPP_INCLUDED
 #define CATCH_SAMPLE_ANALYSIS_HPP_INCLUDED
 
-#include <catch2/benchmark/catch_clock.hpp>
 #include <catch2/benchmark/catch_estimate.hpp>
 #include <catch2/benchmark/catch_outlier_classification.hpp>
 #include <catch2/internal/catch_move_and_forward.hpp>
 
-#include <algorithm>
 #include <vector>
-#include <iterator>
 
 namespace Catch {
     namespace Benchmark {
@@ -33,7 +30,9 @@ namespace Catch {
             operator SampleAnalysis<Duration2>() const {
                 std::vector<Duration2> samples2;
                 samples2.reserve(samples.size());
-                std::transform(samples.begin(), samples.end(), std::back_inserter(samples2), [](Duration d) { return Duration2(d); });
+                for (auto const& d : samples) {
+                    samples2.push_back(Duration2(d));
+                }
                 return {
                     CATCH_MOVE(samples2),
                     mean,
diff --git a/src/catch2/benchmark/detail/catch_analyse.hpp b/src/catch2/benchmark/detail/catch_analyse.hpp
index 1751543c..c932ff26 100644
--- a/src/catch2/benchmark/detail/catch_analyse.hpp
+++ b/src/catch2/benchmark/detail/catch_analyse.hpp
@@ -16,8 +16,6 @@
 #include <catch2/interfaces/catch_interfaces_config.hpp>
 #include <catch2/internal/catch_move_and_forward.hpp>
 
-#include <algorithm>
-#include <iterator>
 #include <vector>
 
 namespace Catch {
@@ -28,7 +26,9 @@ namespace Catch {
                 if (!cfg.benchmarkNoAnalysis()) {
                     std::vector<double> samples;
                     samples.reserve(static_cast<size_t>(last - first));
-                    std::transform(first, last, std::back_inserter(samples), [](Duration d) { return d.count(); });
+                    for (auto current = first; current != last; ++current) {
+                        samples.push_back( current->count() );
+                    }
 
                     auto analysis = Catch::Benchmark::Detail::analyse_samples(cfg.benchmarkConfidenceInterval(), cfg.benchmarkResamples(), samples.begin(), samples.end());
                     auto outliers = Catch::Benchmark::Detail::classify_outliers(samples.begin(), samples.end());
@@ -43,7 +43,10 @@ namespace Catch {
                     };
                     std::vector<Duration> samples2;
                     samples2.reserve(samples.size());
-                    std::transform(samples.begin(), samples.end(), std::back_inserter(samples2), [](double d) { return Duration(d); });
+                    for (auto s : samples) {
+                        samples2.push_back( Duration( s ) );
+                    }
+
                     return {
                         CATCH_MOVE(samples2),
                         wrap_estimate(analysis.mean),
diff --git a/src/catch2/benchmark/detail/catch_estimate_clock.hpp b/src/catch2/benchmark/detail/catch_estimate_clock.hpp
index fb65450e..1e916ae4 100644
--- a/src/catch2/benchmark/detail/catch_estimate_clock.hpp
+++ b/src/catch2/benchmark/detail/catch_estimate_clock.hpp
@@ -19,7 +19,6 @@
 #include <catch2/internal/catch_unique_ptr.hpp>
 
 #include <algorithm>
-#include <iterator>
 #include <vector>
 #include <cmath>
 
@@ -30,13 +29,16 @@ namespace Catch {
             std::vector<double> resolution(int k) {
                 std::vector<TimePoint<Clock>> times;
                 times.reserve(static_cast<size_t>(k + 1));
-                std::generate_n(std::back_inserter(times), k + 1, now<Clock>{});
+                for ( int i = 0; i < k + 1; ++i ) {
+                    times.push_back( Clock::now() );
+                }
 
                 std::vector<double> deltas;
                 deltas.reserve(static_cast<size_t>(k));
-                std::transform(std::next(times.begin()), times.end(), times.begin(),
-                    std::back_inserter(deltas),
-                    [](TimePoint<Clock> a, TimePoint<Clock> b) { return static_cast<double>((a - b).count()); });
+                for ( size_t idx = 1; idx < times.size(); ++idx ) {
+                    deltas.push_back( static_cast<double>(
+                        ( times[idx] - times[idx - 1] ).count() ) );
+                }
 
                 return deltas;
             }
@@ -84,9 +86,11 @@ namespace Catch {
                 std::vector<double> times;
                 int nsamples = static_cast<int>(std::ceil(time_limit / r.elapsed));
                 times.reserve(static_cast<size_t>(nsamples));
-                std::generate_n(std::back_inserter(times), nsamples, [time_clock, &r] {
-                    return static_cast<double>((time_clock(r.iterations) / r.iterations).count());
-                });
+                for ( int s = 0; s < nsamples; ++s ) {
+                    times.push_back( static_cast<double>(
+                        ( time_clock( r.iterations ) / r.iterations )
+                            .count() ) );
+                }
                 return {
                     FloatDuration<Clock>(mean(times.begin(), times.end())),
                     classify_outliers(times.begin(), times.end()),
diff --git a/src/catch2/benchmark/detail/catch_stats.cpp b/src/catch2/benchmark/detail/catch_stats.cpp
index c6c87be3..a7863224 100644
--- a/src/catch2/benchmark/detail/catch_stats.cpp
+++ b/src/catch2/benchmark/detail/catch_stats.cpp
@@ -13,7 +13,7 @@
 
 #include <cassert>
 #include <cstddef>
-#include <iterator>
+#include <numeric>
 #include <random>
 
 
@@ -35,12 +35,19 @@ using Catch::Benchmark::Detail::sample;
 
          sample out;
          out.reserve(resamples);
-         std::generate_n(std::back_inserter(out), resamples, [n, first, &estimator, &dist, &rng] {
-             std::vector<double> resampled;
-             resampled.reserve(n);
-             std::generate_n(std::back_inserter(resampled), n, [first, &dist, &rng] { return first[static_cast<std::ptrdiff_t>(dist(rng))]; });
-             return estimator(resampled.begin(), resampled.end());
-         });
+         // We allocate the vector outside the loop to avoid realloc per resample
+         std::vector<double> resampled;
+         resampled.reserve( n );
+         for ( size_t i = 0; i < resamples; ++i ) {
+             resampled.clear();
+             for ( size_t s = 0; s < n; ++s ) {
+                 resampled.push_back(
+                     first[static_cast<std::ptrdiff_t>( dist( rng ) )] );
+             }
+             const auto estimate =
+                 estimator( resampled.begin(), resampled.end() );
+             out.push_back( estimate );
+         }
          std::sort(out.begin(), out.end());
          return out;
      }
diff --git a/src/catch2/benchmark/detail/catch_stats.hpp b/src/catch2/benchmark/detail/catch_stats.hpp
index b71c5842..62ca80f2 100644
--- a/src/catch2/benchmark/detail/catch_stats.hpp
+++ b/src/catch2/benchmark/detail/catch_stats.hpp
@@ -15,8 +15,6 @@
 
 #include <algorithm>
 #include <vector>
-#include <numeric>
-#include <tuple>
 #include <cmath>
 
 namespace Catch {
@@ -77,13 +75,13 @@ namespace Catch {
 
                 sample jack = jackknife(estimator, first, last);
                 double jack_mean = mean(jack.begin(), jack.end());
-                double sum_squares, sum_cubes;
-                std::tie(sum_squares, sum_cubes) = std::accumulate(jack.begin(), jack.end(), std::make_pair(0., 0.), [jack_mean](std::pair<double, double> sqcb, double x) -> std::pair<double, double> {
-                    auto d = jack_mean - x;
-                    auto d2 = d * d;
-                    auto d3 = d2 * d;
-                    return { sqcb.first + d2, sqcb.second + d3 };
-                });
+                double sum_squares = 0, sum_cubes = 0;
+                for (double x : jack) {
+                    auto difference = jack_mean - x;
+                    auto square = difference * difference;
+                    auto cube = square * difference;
+                    sum_squares += square; sum_cubes += cube;
+                }
 
                 double accel = sum_cubes / (6 * std::pow(sum_squares, 1.5));
                 long n = static_cast<long>(resample.size());