[libcxx-commits] [libcxx] [libc++] Add remaining benchmarks from [alg.modifying.operations] (PR #127354)

[Louis Dionne via libcxx-commits]

================
@@ -0,0 +1,71 @@
+//===----------------------------------------------------------------------===//
+//
+// 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
+//
+//===----------------------------------------------------------------------===//
+
+// UNSUPPORTED: c++03, c++11, c++14, c++17
+
+#include <algorithm>
+#include <cstddef>
+#include <deque>
+#include <iterator>
+#include <list>
+#include <string>
+#include <vector>
+
+#include "benchmark/benchmark.h"
+#include "../../GenerateInput.h"
+
+int main(int argc, char** argv) {
+  auto std_transform = [](auto first1, auto last1, auto first2, auto, auto out, auto f) {
+    return std::transform(first1, last1, first2, out, f);
+  };
+
+  // {std,ranges}::transform(normal container, normal container)
+  {
+    auto bm = []<class Container>(std::string name, auto transform) {
+      benchmark::RegisterBenchmark(
+          name,
+          [transform](auto& st) {
+            std::size_t const size = st.range(0);
+            using ValueType        = typename Container::value_type;
+            Container c1, c2;
+            std::generate_n(std::back_inserter(c1), size, [] { return Generate<ValueType>::random(); });
+            std::generate_n(std::back_inserter(c2), size, [] { return Generate<ValueType>::random(); });
+
+            std::vector<ValueType> out(size);
+
+            auto f = [](auto& x, auto& y) {
+              benchmark::DoNotOptimize(x);
+              benchmark::DoNotOptimize(y);
+              return x + y;
+            };
+
+            for ([[maybe_unused]] auto _ : st) {
+              benchmark::DoNotOptimize(c1);
+              benchmark::DoNotOptimize(c2);
+              benchmark::DoNotOptimize(out);
+              auto result = transform(c1.begin(), c1.end(), c2.begin(), c2.end(), out.begin(), f);
+              benchmark::DoNotOptimize(result);
+            }
+          })
+          ->Arg(32)
+          ->Arg(1024)
+          ->Arg(8192);
+    };
+    bm.operator()<std::vector<int>>("std::transform(vector<int>, vector<int>)", std_transform);
----------------
ldionne wrote:

I agree, but after consideration I would like to tackle that as a separate patch. Indeed, adding this variation of the benchmark requires writing a new batched benchmark since I can't just alternate between the source and the destination range. I'd like to do that in a separate patch that also tackles the benchmarks for `std::copy` (which have already landed).

https://github.com/llvm/llvm-project/pull/127354