dakingffo · caomengxuan666 · May 3, 2026
diff --git a/.gitignore b/.gitignore
@@ -10,6 +10,7 @@ compile_commands.json
 CTestTestfile.cmake
 _deps
 CMakeUserPresets.json
+out/
 
 # CLion
 #  JetBrains specific template is maintained in a separate JetBrains.gitignore that can

diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -17,8 +17,20 @@ FetchContent_Declare(
     FIND_PACKAGE_ARGS NAMES benchmark CONFIG
 )
 set(BENCHMARK_ENABLE_TESTING OFF)
+set(BENCHMARK_ENABLE_WERROR OFF CACHE BOOL "" FORCE)
 FetchContent_MakeAvailable(googlebenchmark)
 
+if(TARGET benchmark)
+    target_compile_options(benchmark PRIVATE
+        $<$<CXX_COMPILER_ID:Clang>:-Wno-c2y-extensions -Wno-invalid-offsetof>
+    )
+endif()
+if(TARGET benchmark_main)
+    target_compile_options(benchmark_main PRIVATE
+        $<$<CXX_COMPILER_ID:Clang>:-Wno-c2y-extensions -Wno-invalid-offsetof>
+    )
+endif()
+
 # Google Test
 enable_testing()
 FetchContent_Declare(
@@ -39,8 +51,19 @@ set(HDR_HISTOGRAM_BUILD_SHARED OFF CACHE BOOL "" FORCE)
 set(HDR_LOG_REQUIRED "DISABLED" CACHE STRING "" FORCE)
 FetchContent_MakeAvailable(hdrhistogram)
 
+# moodycamel ConcurrentQueue
+FetchContent_Declare(
+    concurrentqueue
+    URL https://github.com/cameron314/concurrentqueue/archive/refs/tags/v1.0.5.zip
+)
+FetchContent_MakeAvailable(concurrentqueue)
+
 find_package(Threads REQUIRED)
 
+set(ATOMIC_LIBRARY
+    $<$<AND:$<NOT:$<CXX_COMPILER_ID:MSVC>>,$<NOT:$<PLATFORM_ID:Windows>>>:atomic>
+)
+
 set(COMMON_TARGET_PROPERTIES
     PRIVATE
         ${DEFAULT_RELEASE_OPTS}
@@ -58,7 +81,7 @@ target_link_libraries(mpsc_bench_throughput
     PRIVATE
         benchmark::benchmark_main
         Threads::Threads
-        $<$<NOT:$<CXX_COMPILER_ID:MSVC>>:atomic>
+        ${ATOMIC_LIBRARY}
 )
 target_compile_options(mpsc_bench_throughput ${COMMON_TARGET_PROPERTIES})
 
@@ -74,7 +97,7 @@ target_link_libraries(mpsc_bench_latency
         benchmark::benchmark_main
         hdr_histogram_static 
         Threads::Threads
-        $<$<NOT:$<CXX_COMPILER_ID:MSVC>>:atomic>
+        ${ATOMIC_LIBRARY}
 )
 target_compile_options(mpsc_bench_latency ${COMMON_TARGET_PROPERTIES})
 
@@ -88,24 +111,39 @@ target_link_libraries(mpsc_bench_linearizable
     PRIVATE
         benchmark::benchmark_main
         Threads::Threads
-        $<$<NOT:$<CXX_COMPILER_ID:MSVC>>:atomic>
+        ${ATOMIC_LIBRARY}
 )
 target_compile_options(mpsc_bench_linearizable ${COMMON_TARGET_PROPERTIES})
 
+add_executable(mpsc_vs_mpmc_benchmark benchmarks/bench_mpsc_vs_mpmc_google.cpp)
+target_include_directories(mpsc_vs_mpmc_benchmark
+    PRIVATE
+        ${CMAKE_CURRENT_SOURCE_DIR}/include
+        ${CMAKE_CURRENT_SOURCE_DIR}
+        ${concurrentqueue_SOURCE_DIR}
+)
+target_link_libraries(mpsc_vs_mpmc_benchmark
+    PRIVATE
+        benchmark::benchmark_main
+        Threads::Threads
+        ${ATOMIC_LIBRARY}
+)
+target_compile_options(mpsc_vs_mpmc_benchmark ${COMMON_TARGET_PROPERTIES})
+
 # TEST
 add_executable(mpsc_tests tests/test_MPSC.cpp)
 target_include_directories(mpsc_tests PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/include ${CMAKE_CURRENT_SOURCE_DIR})
-target_link_libraries(mpsc_tests PRIVATE GTest::gtest_main GTest::gmock Threads::Threads $<$<NOT:$<CXX_COMPILER_ID:MSVC>>:atomic>)
+target_link_libraries(mpsc_tests PRIVATE GTest::gtest_main GTest::gmock Threads::Threads ${ATOMIC_LIBRARY})
 include(GoogleTest)
 gtest_discover_tests(mpsc_tests)
 
 #EXAMPLE
 add_executable(mpsc_log_system_example examples/log_system.cpp)
 target_include_directories(mpsc_log_system_example PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/include ${CMAKE_CURRENT_SOURCE_DIR})
-target_link_libraries(mpsc_log_system_example PRIVATE Threads::Threads $<$<NOT:$<CXX_COMPILER_ID:MSVC>>:atomic>)
+target_link_libraries(mpsc_log_system_example PRIVATE Threads::Threads ${ATOMIC_LIBRARY})
 target_compile_options(mpsc_log_system_example ${COMMON_TARGET_PROPERTIES})
 
 add_executable(mpsc_command_dispatcher_example examples/command_dispatcher.cpp)
 target_include_directories(mpsc_command_dispatcher_example PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}/include ${CMAKE_CURRENT_SOURCE_DIR})
-target_link_libraries(mpsc_command_dispatcher_example PRIVATE Threads::Threads $<$<NOT:$<CXX_COMPILER_ID:MSVC>>:atomic>)
-target_compile_options(mpsc_command_dispatcher_example ${COMMON_TARGET_PROPERTIES})
+target_link_libraries(mpsc_command_dispatcher_example PRIVATE Threads::Threads ${ATOMIC_LIBRARY})
+target_compile_options(mpsc_command_dispatcher_example ${COMMON_TARGET_PROPERTIES})
diff --git a/README.md b/README.md
@@ -186,8 +186,64 @@ This scenario tests the throughput when producers use the `enqueue_bulk` interfa
 | **moodycamel** | **2** | **1** | **68.9161** | **Best performance in this section (P=2)** | |
 | moodycamel | 16 | 1 | 43.4848 | No bulk advantage at 16P |
 
+**Part IV: Additional Google Benchmark Comparison (Clang Release)**
 
-**Part IV: Enqueue/Dequeue Latency**
+This comparison is produced by `mpsc_vs_mpmc_benchmark`, using Google Benchmark in a Clang Release build. It uses `moodycamel::ConcurrentQueue` v1.0.5 as a generic MPMC baseline under the same MPSC workloads.
+
+Run on (20 X 2688 MHz CPU s)
+CPU Caches:
+L1 Data 48 KiB (x10)
+L1 Instruction 32 KiB (x10)
+L2 Unified 1280 KiB (x10)
+L3 Unified 24576 KiB (x1)
+Compiler: Clang 22.1.1 Release
+
+Command:
+
+```powershell
+cmake -S . -B out/build/clang-local -G Ninja -DCMAKE_CXX_COMPILER=clang++ -DCMAKE_BUILD_TYPE=Release
+cmake --build out/build/clang-local --target mpsc_vs_mpmc_benchmark
+.\out\build\clang-local\mpsc_vs_mpmc_benchmark.exe --benchmark_min_time=0.1s --benchmark_repetitions=1 --benchmark_counters_tabular=true
+```
+
+Uniform single-element enqueue:
+
+| Queue | P (Producers) | C (Consumer) | **Throughput (M items/s)** |
+| :--- | :--- | :--- | :--- |
+| **daking** | 1 | 1 | **86.78** |
+| daking | 2 | 1 | **32.29** |
+| daking | 4 | 1 | **32.52** |
+| daking | 8 | 1 | 31.16 |
+| moodycamel | 1 | 1 | 22.74 |
+| moodycamel | 2 | 1 | 27.82 |
+| moodycamel | 4 | 1 | 29.79 |
+| **moodycamel** | 8 | 1 | **32.07** |
+
+Uneven sequential burst:
+
+| Queue | P (Producers) | C (Consumer) | Relay % | **Throughput (M items/s)** |
+| :--- | :--- | :--- | :--- | :--- |
+| daking | 4 | 1 | $50.0\%$ | **33.89** |
+| daking | 4 | 1 | $90.0\%$ | **61.72** |
+| daking | 4 | 1 | $98.0\%$ | **73.72** |
+| moodycamel | 4 | 1 | $50.0\%$ | 23.84 |
+| moodycamel | 4 | 1 | $90.0\%$ | 22.25 |
+| moodycamel | 4 | 1 | $98.0\%$ | 17.64 |
+
+Bulk enqueue:
+
+| Queue | P (Producers) | C (Consumer) | **Throughput (M items/s)** |
+| :--- | :--- | :--- | :--- |
+| **daking** | 1 | 1 | **102.25** |
+| **daking** | 2 | 1 | **102.37** |
+| **daking** | 4 | 1 | **85.23** |
+| **daking** | 8 | 1 | **77.16** |
+| moodycamel | 1 | 1 | 17.02 |
+| moodycamel | 2 | 1 | 18.85 |
+| moodycamel | 4 | 1 | 18.06 |
+| moodycamel | 8 | 1 | 16.02 |
+
+**Part V: Enqueue/Dequeue Latency**
 
 (Based on HdrHistogram, Test on Linux)
 We get below performance：

diff --git a/README.zh.md b/README.zh.md
@@ -186,7 +186,64 @@ Compiler: MSVC -O2
 | **moodycamel** | **2** | **1** | **68.9161** | **本部分最高表现 (P=2)** |
 | moodycamel | 16 | 1 | 43.4848 | 16P时无明显批量优势 |
 
-**第四部分：Enqueue/Dequeue Latency**
+**第四部分：额外 Google Benchmark 对比（Clang Release）**
+
+此对比由 `mpsc_vs_mpmc_benchmark` 产生，使用 Google Benchmark 和 Clang Release 构建。这里将 `moodycamel::ConcurrentQueue` v1.0.5 作为通用 MPMC 基线，并在相同 MPSC 负载下进行对比。
+
+Run on (20 X 2688 MHz CPU s)
+CPU Caches:
+L1 Data 48 KiB (x10)
+L1 Instruction 32 KiB (x10)
+L2 Unified 1280 KiB (x10)
+L3 Unified 24576 KiB (x1)
+Compiler: Clang 22.1.1 Release
+
+运行命令：
+
+```powershell
+cmake -S . -B out/build/clang-local -G Ninja -DCMAKE_CXX_COMPILER=clang++ -DCMAKE_BUILD_TYPE=Release
+cmake --build out/build/clang-local --target mpsc_vs_mpmc_benchmark
+.\out\build\clang-local\mpsc_vs_mpmc_benchmark.exe --benchmark_min_time=0.1s --benchmark_repetitions=1 --benchmark_counters_tabular=true
+```
+
+均匀单元素入队：
+
+| 队列 | P (生产者) | C (消费者) | **吞吐量 (M items/s)** |
+| :--- | :--- | :--- | :--- |
+| **daking** | 1 | 1 | **86.78** |
+| daking | 2 | 1 | **32.29** |
+| daking | 4 | 1 | **32.52** |
+| daking | 8 | 1 | 31.16 |
+| moodycamel | 1 | 1 | 22.74 |
+| moodycamel | 2 | 1 | 27.82 |
+| moodycamel | 4 | 1 | 29.79 |
+| **moodycamel** | 8 | 1 | **32.07** |
+
+不均匀顺序爆发：
+
+| 队列 | P (生产者) | C (消费者) | 接力百分比 | **吞吐量 (M items/s)** |
+| :--- | :--- | :--- | :--- | :--- |
+| daking | 4 | 1 | $50.0\%$ | **33.89** |
+| daking | 4 | 1 | $90.0\%$ | **61.72** |
+| daking | 4 | 1 | $98.0\%$ | **73.72** |
+| moodycamel | 4 | 1 | $50.0\%$ | 23.84 |
+| moodycamel | 4 | 1 | $90.0\%$ | 22.25 |
+| moodycamel | 4 | 1 | $98.0\%$ | 17.64 |
+
+批量入队：
+
+| 队列 | P (生产者) | C (消费者) | **吞吐量 (M items/s)** |
+| :--- | :--- | :--- | :--- |
+| **daking** | 1 | 1 | **102.25** |
+| **daking** | 2 | 1 | **102.37** |
+| **daking** | 4 | 1 | **85.23** |
+| **daking** | 8 | 1 | **77.16** |
+| moodycamel | 1 | 1 | 17.02 |
+| moodycamel | 2 | 1 | 18.85 |
+| moodycamel | 4 | 1 | 18.06 |
+| moodycamel | 8 | 1 | 16.02 |
+
+**第五部分：Enqueue/Dequeue Latency**
 
 (此部分基于HdrHistogram，在Linux平台测试)
 我们得到了以下延迟表现：
@@ -323,4 +380,4 @@ daking::MPSC_queue<double> queue3;
 
 ## 许可证 (LICENSE)
 
-MPSC\_queue 使用 [MIT 许可证](https://www.google.com/search?q=./LICENSE.txt) 授权。
+MPSC\_queue 使用 [MIT 许可证](https://www.google.com/search?q=./LICENSE.txt) 授权。
-Original file line number
+Diff line change
@@ Expand Up / @@ -10,6 +10,7 @@ compile_commands.json @@
     CTestTestfile.cmake
     _deps
     CMakeUserPresets.json
+    out/
     # CLion
     #  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
@@ Expand Down @@