mirror of
https://github.com/Light-City/CPlusPlusThings.git
synced 2026-10-29 03:14:06 +08:00
dfe1c933b3
- const README: rewrite the space-saving claim per issue #5 follow-ups (constant-expression const vs #define immediates; which const qualify as constant expressions) - const README: fix the class-section diagnosis per issue #338 (overload resolution picks the const overload; the restriction only surfaces once that overload is removed) - sizeof README: virtual-inheritance result is compiler-dependent (GCC merges vbptr+vfptr -> 16, MSVC keeps them separate -> 32) #246 - static README: new section on internal linkage / access restriction with anonymous-namespace comparison #142 - 10-day practice: full problem list table + fixed day links #136 - new design_pattern/singleton/singleton_default.hpp: reusable thread-safe CRTP singleton template, compile+concurrency tested #303
54 lines
1.8 KiB
Markdown
54 lines
1.8 KiB
Markdown
> 来自 [C++那些事网站](https://light-city.github.io/stories_things/cpp-concurrency/false-sharing-cache-line/) 的文章，由站点同步脚本导出。
|
||
|
||
# C++ False Sharing 与缓存行
|
||
|
||
最近看到一段代码，手动做的对齐，于是研究一下不对齐又会带来什么影响？
|
||
|
||
```cpp
|
||
template <typename T>
|
||
class AtomicWithPadding {
|
||
private:
|
||
static constexpr int kCacheLineSize = 64;
|
||
uint8_t padding_before[kCacheLineSize];
|
||
|
||
public:
|
||
std::atomic<T> value;
|
||
|
||
private:
|
||
uint8_t padding_after[kCacheLineSize];
|
||
};
|
||
```
|
||
|
||
这里便引出了大家面试的必考点：
|
||
|
||
- Cache line
|
||
|
||
于是有下面几个面试问题，你能答上来？
|
||
|
||
- 什么是cache line?
|
||
- 什么是MESI协议？
|
||
- 什么是False sharing?
|
||
- C++里面如何保证避免False sharing?
|
||
- 避免False sharing可以带来什么好处？
|
||
- 你知道C++17之后提供了哪两个变量解决这个问题？
|
||
|
||
于是，找了个性能测试，大概实现是：
|
||
|
||
- 定义4个变量，a、b、c、d，多线程进行共享
|
||
- 另外是对这四个变量对齐，然后开n个线程去调度。
|
||
|
||
下面是benchmark：
|
||
|
||
```cpp
|
||
------------------------------------------------------------------
|
||
Benchmark Time CPU Iterations
|
||
------------------------------------------------------------------
|
||
singleThread 4.05 ms 4.04 ms 133
|
||
directSharing/real_time 8.78 ms 0.069 ms 77
|
||
falseSharing/real_time 8.81 ms 0.083 ms 74
|
||
noSharing/real_time 1.16 ms 0.071 ms 578
|
||
```
|
||
|
||
可以看到noSharing得到了非常好的效果！缓存命中率大大提高，当然这里用perf分析一下是更好的，可以看看L1 cache miss。
|
||
|
||
以上答案在知识星球已更新，欢迎一起探讨，留言区交流，谢谢~ |