Windows .NET (C++) 通过 YIELD/PAUSE 指令为超线程友好的忙等待提供 SpinWait。Linux 和 Mac OS X 上的等效功能是什么?如果系统调用不可用,如何在用户空间实现等价的?
有关自旋等待的性能问题的讨论,请参阅启用超线程技术的英特尔处理器上的长时间自旋等待循环。
Windows .NET (C++) 通过 YIELD/PAUSE 指令为超线程友好的忙等待提供 SpinWait。Linux 和 Mac OS X 上的等效功能是什么?如果系统调用不可用,如何在用户空间实现等价的?
有关自旋等待的性能问题的讨论,请参阅启用超线程技术的英特尔处理器上的长时间自旋等待循环。
参见sameer_87的https://www.codeproject.com/articles/184046/spin-lock-in-c
#include "SpinLock.h"
#include <iostream>
using namespace LockFree;
using namespace std;
void tSpinWait::Lock(tSpinLock &LockObj)
{
m_iterations = 0;
while(true)
{
// A thread alreading owning the lock shouldn't be allowed to wait to acquire the lock - reentrant safe
if(LockObj.dest == GetCurrentThreadId())
break;
/*
Spinning in a loop of interlockedxxx calls can reduce the available memory bandwidth and slow
down the rest of the system. Interlocked calls are expensive in their use of the system memory
bus. It is better to see if the 'dest' value is what it is expected and then retry interlockedxx.
*/
if(InterlockedCompareExchange(&LockObj.dest, LockObj.exchange, LockObj.compare) == 0)
{
//assign CurrentThreadId to dest to make it re-entrant safe
LockObj.dest = GetCurrentThreadId();
// lock acquired
break;
}
// spin wait to acquire
while(LockObj.dest != LockObj.compare)
{
if(HasThreasholdReached())
{
if(m_iterations + YIELD_ITERATION >= MAX_SLEEP_ITERATION)
Sleep(0);
if(m_iterations >= YIELD_ITERATION && m_iterations < MAX_SLEEP_ITERATION)
{
m_iterations = 0;
SwitchToThread();
}
}
// Yield processor on multi-processor but if on single processor then give other thread the CPU
m_iterations++;
if(Helper::GetNumberOfProcessors() > 1) { YieldProcessor(/*no op*/); }
else { SwitchToThread(); }
}
}
}
//
void tSpinWait::Unlock(tSpinLock &LockObj)
{
if(LockObj.dest != GetCurrentThreadId())
throw std::runtime_error("Unexpected thread-id in release");
// lock released
InterlockedCompareExchange(&LockObj.dest, LockObj.compare, GetCurrentThreadId());
}