[pull] master from parallel101:master

08/include/CudaAllocator.h

@@ -25,6 +25,10 @@ struct CudaAllocator {
             ::new((void *)p) T(std::forward<Args>(args)...);
     }
 
+    template <class U> // cihou shabi wendous
+    constexpr CudaAllocator(CudaAllocator<U> const &other) noexcept {
+    }
+
     constexpr bool operator==(CudaAllocator<T> const &other) const {
         return this == &other;
     }

09/include/CudaAllocator.h

@@ -25,6 +25,10 @@ struct CudaAllocator {
             ::new((void *)p) T(std::forward<Args>(args)...);
     }
 
+    template <class U> // cihou shabi wendous
+    constexpr CudaAllocator(CudaAllocator<U> const &other) noexcept {
+    }
+
     constexpr bool operator==(CudaAllocator<T> const &other) const {
         return this == &other;
     }

specifelse/benchtest/.gitignore

@@ -0,0 +1,2 @@
+jomtSettings/
+.cache/

specifelse/benchtest/.tasks

@@ -0,0 +1,4 @@
+[+]
+build_type=Release
+build_target=demo0
+run_target="$(VIM:build_dir)/$(VIM:build_target)" --benchmark_out="$(VIM:build_dir)/$(VIM:build_target).json"

specifelse/benchtest/CMakeLists.txt

@@ -0,0 +1,18 @@
+cmake_minimum_required(VERSION 3.15)
+
+if (NOT CMAKE_BUILD_TYPE)
+    set(CMAKE_BUILD_TYPE Release)
+endif()
+set(CMAKE_CXX_STANDARD 17)
+
+project(BenchCourse LANGUAGES CXX)
+add_subdirectory(benchmark)
+
+set(DEMO_LIST
+    demo0
+)
+
+foreach (name IN ITEMS ${DEMO_LIST})
+    add_executable(${name} ${name}.cpp)
+    target_link_libraries(${name} PRIVATE benchmark)
+endforeach()

specifelse/benchtest/demo0.cpp

@@ -0,0 +1,68 @@
+#include <benchmark/benchmark.h>
+#include <random>
+#include <vector>
+#include <algorithm>
+
+void upperIfelse(char *p, int n) {
+    for (int i = 0; i < n; i++) {
+        if ('a' <= p[i] && p[i] <= 'z')
+            p[i] = p[i] + 'A' - 'a';
+    }
+}
+
+void upperTernary(char *p, int n) {
+    for (int i = 0; i < n; i++) {
+        p[i] = ('a' <= p[i] && p[i] <= 'z') ? p[i] + 'A' - 'a' : p[i];
+    }
+}
+
+template <void (*upper)(char *p, int n)>
+void testSorted(benchmark::State &state) {
+    const int n = (int)1e7;
+    std::vector<char> a(n);
+
+    std::mt19937 rng;
+    std::uniform_int_distribution<char> dist(0, 127);
+    for (int i = 0; i < n; i++) {
+        a[i] = dist(rng);
+    }
+    std::sort(a.begin(), a.end());
+
+    for (auto _: state) {
+        upper(a.data(), n);
+        benchmark::DoNotOptimize(a);
+    }
+}
+
+template <void (*upper)(char *p, int n)>
+void testRandom(benchmark::State &state) {
+    const int n = (int)1e7;
+    std::vector<char> a(n);
+
+    std::mt19937 rng;
+    std::uniform_int_distribution<char> dist(0, 127);
+    for (int i = 0; i < n; i++) {
+        a[i] = dist(rng);
+    }
+
+    for (auto _: state) {
+        upper(a.data(), n);
+        benchmark::DoNotOptimize(a);
+    }
+}
+
+BENCHMARK_TEMPLATE(testSorted, upperTernary);
+BENCHMARK_TEMPLATE(testSorted, upperIfelse);
+BENCHMARK_TEMPLATE(testRandom, upperTernary);
+BENCHMARK_TEMPLATE(testRandom, upperIfelse);
+
+BENCHMARK_MAIN();
+/* 
+-------------------------------------------------------------------
+Benchmark                         Time             CPU   Iterations
+-------------------------------------------------------------------
+testSorted<upperTernary>     547329 ns       546965 ns         1279
+testSorted<upperIfelse>     4597908 ns      4597300 ns          153
+testRandom<upperTernary>     548555 ns       548232 ns         1296
+testRandom<upperIfelse>     7031812 ns      7031073 ns           87
+*/

specifelse/benchtest/jomt/.gitignore

@@ -0,0 +1,35 @@
+# Prerequisites
+*.d
+
+# Compiled Object files
+*.slo
+*.lo
+*.o
+*.obj
+
+# Precompiled Headers
+*.gch
+*.pch
+
+# Compiled Dynamic libraries
+*.so
+*.dylib
+*.dll
+
+# Fortran module files
+*.mod
+*.smod
+
+# Compiled Static libraries
+*.lai
+*.la
+*.a
+*.lib
+
+# Executables
+*.exe
+*.out
+*.app
+
+# Misc
+CMakeLists.txt.user

specifelse/benchtest/jomt/.tasks

@@ -0,0 +1,4 @@
+[+]
+build_type=Release
+build_target=JOMT
+run_target="$(VIM:build_dir)/src/$(VIM:build_target)"

specifelse/benchtest/jomt/CMakeLists.txt

@@ -0,0 +1,23 @@
+cmake_minimum_required(VERSION 3.5)
+
+project(JOMT LANGUAGES CXX)
+
+set(CMAKE_INCLUDE_CURRENT_DIR ON)
+
+set(CMAKE_AUTOUIC ON)
+set(CMAKE_AUTOMOC ON)
+set(CMAKE_AUTORCC ON)
+
+set(CMAKE_CXX_STANDARD 11)
+set(CMAKE_CXX_STANDARD_REQUIRED ON)
+
+if (MSVC)
+  add_compile_options(/W3)
+  set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} /O2 /DNDEBUG /DQT_NO_DEBUG_OUTPUT")
+else()
+  add_compile_options(-Wall -Wextra -pedantic)
+  set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -O2 -DNDEBUG -DQT_NO_DEBUG_OUTPUT -march=native -mtune=native")
+endif()
+
+# Build the target
+add_subdirectory(src)