剖析 C++ 函数性能优化盲区，深入优化见真章

// Using a memory pool
std::pmr::unsynchronized_pool_resource mem_pool;
std::pmr::memory_resource* pool_resource = &mem_pool;
std::pmr::polymorphic_allocator<int> allocator(pool_resource);

// Allocating using the memory pool
int* p = allocator.allocate(100);

// Using move semantics
std::vector<int> vec1;
std::vector<int> vec2;
vec2 = std::move(vec1); // Move semantics avoid copying elements

// Using references to pass arguments to functions
void foo(const std::vector<int>& v) {
    // Use v as read-only without copying
}

// Inline small functions
inline int sum(int a, int b) { return a + b; }

// Use lambda expressions to eliminate function call overhead
std::vector<int> v;
std::sort(v.begin(), v.end(), [](int a, int b) { return a < b; });

// Accessing elements in cache-friendly order
int arr[] = {0, 1, 2, 3, 4, 5, 6, 7};
for (int i = 0; i < N; i += 16) {
    // Access elements in cache-friendly chunks
    for (int j = 0; j < 16; j++)
        arr[i + j] = ...;
}

// Using branch prediction hints
int foo(int n) {
    if (n >= 0)  // Most likely branch
        return n;
    else  // Least likely branch
        return -n;
}