[libc++] Optimize ranges::count for __bit_iterators

```
---------------------------------------------------------------
Benchmark                                    old            new
---------------------------------------------------------------
bm_vector_bool_count/1                   1.92 ns        1.92 ns
bm_vector_bool_count/2                   1.92 ns        1.92 ns
bm_vector_bool_count/3                   1.92 ns        1.92 ns
bm_vector_bool_count/4                   1.92 ns        1.92 ns
bm_vector_bool_count/5                   1.92 ns        1.92 ns
bm_vector_bool_count/6                   1.92 ns        1.92 ns
bm_vector_bool_count/7                   1.92 ns        1.92 ns
bm_vector_bool_count/8                   1.92 ns        1.92 ns
bm_vector_bool_count/16                  1.92 ns        1.92 ns
bm_vector_bool_count/64                  2.24 ns        2.25 ns
bm_vector_bool_count/512                 3.19 ns        3.20 ns
bm_vector_bool_count/4096                14.1 ns        12.3 ns
bm_vector_bool_count/32768               84.0 ns        83.6 ns
bm_vector_bool_count/262144               664 ns         661 ns
bm_vector_bool_count/1048576             2623 ns        2628 ns
bm_vector_bool_ranges_count/1            1.07 ns        1.92 ns
bm_vector_bool_ranges_count/2            1.65 ns        1.92 ns
bm_vector_bool_ranges_count/3            2.27 ns        1.92 ns
bm_vector_bool_ranges_count/4            2.68 ns        1.92 ns
bm_vector_bool_ranges_count/5            3.33 ns        1.92 ns
bm_vector_bool_ranges_count/6            3.99 ns        1.92 ns
bm_vector_bool_ranges_count/7            4.67 ns        1.92 ns
bm_vector_bool_ranges_count/8            5.19 ns        1.92 ns
bm_vector_bool_ranges_count/16           11.1 ns        1.92 ns
bm_vector_bool_ranges_count/64           52.2 ns        2.24 ns
bm_vector_bool_ranges_count/512           452 ns        3.20 ns
bm_vector_bool_ranges_count/4096         3577 ns        12.1 ns
bm_vector_bool_ranges_count/32768       28725 ns        83.7 ns
bm_vector_bool_ranges_count/262144     229676 ns         662 ns
bm_vector_bool_ranges_count/1048576    905574 ns        2625 ns
```

Reviewed By: #libc, ldionne

Spies: arichardson, ldionne, libcxx-commits

Differential Revision: https://reviews.llvm.org/D156956

2 files changed, 36 insertions, 0 deletions

diff --git a/libcxx/benchmarks/CMakeLists.txt b/libcxx/benchmarks/CMakeLists.txt
index 56d22af0..1a4d634 100644
--- a/libcxx/benchmarks/CMakeLists.txt
+++ b/libcxx/benchmarks/CMakeLists.txt
@@ -173,6 +173,7 @@ endfunction()
 #==============================================================================
 set(BENCHMARK_TESTS
     algorithms.partition_point.bench.cpp
+    algorithms/count.bench.cpp
     algorithms/equal.bench.cpp
     algorithms/find.bench.cpp
     algorithms/lower_bound.bench.cpp
diff --git a/libcxx/benchmarks/algorithms/count.bench.cpp b/libcxx/benchmarks/algorithms/count.bench.cpp
new file mode 100644
index 0000000..7370293
--- /dev/null
+++ b/libcxx/benchmarks/algorithms/count.bench.cpp
@@ -0,0 +1,35 @@
+//===----------------------------------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include <algorithm>
+#include <benchmark/benchmark.h>
+#include <cstring>
+#include <random>
+#include <vector>
+
+static void bm_vector_bool_count(benchmark::State& state) {
+  std::vector<bool> vec1(state.range(), false);
+
+  for (auto _ : state) {
+    benchmark::DoNotOptimize(vec1);
+    benchmark::DoNotOptimize(std::count(vec1.begin(), vec1.end(), true));
+  }
+}
+BENCHMARK(bm_vector_bool_count)->DenseRange(1, 8)->Range(16, 1 << 20);
+
+static void bm_vector_bool_ranges_count(benchmark::State& state) {
+  std::vector<bool> vec1(state.range(), false);
+
+  for (auto _ : state) {
+    benchmark::DoNotOptimize(vec1);
+    benchmark::DoNotOptimize(std::ranges::count(vec1.begin(), vec1.end(), true));
+  }
+}
+BENCHMARK(bm_vector_bool_ranges_count)->DenseRange(1, 8)->Range(16, 1 << 20);
+
+BENCHMARK_MAIN();