add simpler benchmarks

bench/simple_bench.cpp

@@ -0,0 +1,131 @@
+#include <pthread.h>
+
+#include <array>
+#include <atomic>
+#include <boost/lockfree/queue.hpp>
+#include <boost/lockfree/spsc_queue.hpp>
+#include <chrono>
+#include <iomanip>
+#include <iostream>
+#include <mutex>
+#include <optional>
+#include <thread>
+#include <vector>
+
+#include "common.hpp"
+#include "mpsc.hpp"
+#include "spsc.hpp"
+
+using namespace fastchan;
+
+constexpr int num_iterations = 100'000'000;
+
+template <typename Chan>
+void producer(Chan &chan, int num_producers, int producer_id) {
+    for (int i = producer_id; i <= num_iterations; i += num_producers) {
+        chan.put(i);
+    }
+}
+
+template <typename Chan>
+void consumer(Chan &chan) {
+    int consumed = 0;
+    while (consumed < num_iterations) {
+        auto item = chan.get();
+        consumed++;
+    }
+}
+
+void set_affinity(std::thread &t, int core_id) {
+    cpu_set_t cpuset;
+    CPU_ZERO(&cpuset);
+    CPU_SET(core_id, &cpuset);
+
+    auto native_handle = t.native_handle();
+    int result = pthread_setaffinity_np(native_handle, sizeof(cpu_set_t), &cpuset);
+    if (result != 0) {
+        std::cerr << "Error setting thread affinity: " << result << std::endl;
+    }
+}
+
+template <typename Chan>
+void run_benchmark(const std::string &name, int num_producers, bool pin_threads) {
+    Chan chan;
+
+    auto start = std::chrono::steady_clock::now();
+
+    std::vector<std::thread> producers;
+    for (int i = 0; i < num_producers; i++) {
+        producers.emplace_back(producer<Chan>, std::ref(chan), num_producers, i);
+        if (pin_threads) {
+            set_affinity(producers.back(), i % std::thread::hardware_concurrency());
+        }
+    }
+    std::thread consumer_thread(consumer<Chan>, std::ref(chan));
+    if (pin_threads) {
+        set_affinity(consumer_thread, num_producers % std::thread::hardware_concurrency());
+    }
+
+    for (auto &t : producers) {
+        t.join();
+    }
+    consumer_thread.join();
+
+    auto end = std::chrono::steady_clock::now();
+    auto elapsed = std::chrono::duration_cast<std::chrono::nanoseconds>(end - start).count();
+    std::cout << name << "\t" << num_producers << " Producers"
+              << "\t" << (pin_threads ? "  Pinned" : "UnPinned") << "\t" << num_iterations << "\t" << std::setfill(' ') << std::setw(3)
+              << (elapsed / num_iterations) << " ns/op"
+              << "\t" << elapsed / 1'000'000 << " ms" << std::endl;
+}
+
+int main() {
+#if defined(__linux__)
+    run_benchmark<SPSC<int, BlockingPutBlockingGet, 1024, WaitYield>>("SPSC_Yield", 1, true);
+    run_benchmark<SPSC<int, BlockingPutBlockingGet, 1024, WaitYield>>("SPSC_Yield", 1, false);
+    run_benchmark<SPSC<int, BlockingPutBlockingGet, 1024, WaitSpin>>("SPSC_Spin", 1, true);
+    run_benchmark<SPSC<int, BlockingPutBlockingGet, 1024, WaitSpin>>("SPSC_Spin", 1, false);
+    run_benchmark<SPSC<int, BlockingPutBlockingGet, 1024, WaitCondition>>("SPSC_Cond", 1, true);
+    run_benchmark<SPSC<int, BlockingPutBlockingGet, 1024, WaitCondition>>("SPSC_Cond", 1, false);
+
+    std::cout << "============================" << std::endl;
+
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitYield>>("MPSC_Yield", 1, true);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitYield>>("MPSC_Yield", 1, false);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitSpin>>("MPSC_Spin", 1, true);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitSpin>>("MPSC_Spin", 1, false);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitCondition>>("MPSC_Cond", 1, true);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitCondition>>("MPSC_Cond", 1, false);
+
+    std::cout << "============================" << std::endl;
+
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitYield>>("MPSC_Yield", 3, true);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitYield>>("MPSC_Yield", 3, false);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitSpin>>("MPSC_Spin", 3, true);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitSpin>>("MPSC_Spin", 3, false);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitCondition>>("MPSC_Cond", 3, true);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitCondition>>("MPSC_Cond", 3, false);
+
+    std::cout << "============================" << std::endl;
+
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitYield>>("MPSC_Yield", 5, true);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitYield>>("MPSC_Yield", 5, false);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitSpin>>("MPSC_Spin", 5, true);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitSpin>>("MPSC_Spin", 5, false);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitCondition>>("MPSC_Cond", 5, true);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitCondition>>("MPSC_Cond", 5, false);
+
+    std::cout << "============================" << std::endl;
+
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitYield>>("MPSC_Yield", 7, true);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitYield>>("MPSC_Yield", 7, false);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitSpin>>("MPSC_Spin", 7, true);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitSpin>>("MPSC_Spin", 7, false);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitCondition>>("MPSC_Cond", 7, true);
+    run_benchmark<MPSC<int, BlockingPutBlockingGet, 1024, WaitCondition>>("MPSC_Cond", 7, false);
+
+#else
+    std::cout << "This benchmark requires a Linux platform to run. Exiting." << std::endl;
+#endif
+    return 0;
+}