2

Using VC++ 13 under windows, the on-line help states that using Sleep(0) relinquishes the remainder of the current threads time slice to any other thread of equal priority. Is this also the case for other values? e.g. if I use Sleep(1000) are 1000ms of CPU time for the core on which the current thread is running likely to be usable by another thread? I imagine this is hardware and implementation specific, so to narrow it assume Intel I5 or better, Windows 7 or 8.

The reason for asking is I have a thread pool class, and I'm using an additional monitor thread to report progress, allow the user to abort long processes, etc...

SmacL
  • 22,555
  • 12
  • 95
  • 149

2 Answers2

3

Yes, zero has the special meaning only in the regard to signal there is no minimal time to wait. Normally it could be interpreted like "I want to sleep for no-time" which doesn't make much sense. It means "I want to give chance to other thread to run."

If it's non-zero, thread is guaranteed not to be returned to for the amount of time specified, of course within the clock resolution. When thread gets suspended it gets a suspended status in the system and is not considered during scheduling. With 0 it doesn't change it's status, so it remains ready to run, and the function might return immediately.

Also, I don't think it is hardware related, this is purely system level thing.

luk32
  • 15,812
  • 38
  • 62
  • 1
    It's certainly not hardware related. This stuff worked on 486'es running NT 3.51. – MSalters Apr 09 '15 at 10:32
  • I am just not sure if there are any processor with their own thread scheduling, I don't think so, but I don't know everything. If there are, some OSes might utilize it, but I don't think I have ever heard of such thing. – luk32 Apr 09 '15 at 10:35
  • Thanks for that, I'm guessing the only hardware related aspect then is the shortest sleep time that may be usable by other threads, i.e. minimum time slice – SmacL Apr 09 '15 at 10:59
  • 1
    @ShaneMacLaughlin It's tied to the system clock, and it usually by default runs at lower resolution than supported by hardware. So the OS abstracts even that. In other words, it could be limited by hardware on relatively slow processors, or some architectures where some low resolution clocks are used: https://msdn.microsoft.com/en-us/library/windows/hardware/dn265247(v=vs.85).aspx . – luk32 Apr 09 '15 at 11:30
1

MSDN: Sleep function

A value of zero causes the thread to relinquish the remainder of its time slice to any other thread that is ready to run. If there are no other threads ready to run, the function returns immediately, and the thread continues execution.

The special XP case is described as follows:

Windows XP: A value of zero causes the thread to relinquish the remainder of its time slice to any other thread of equal priority that is ready to run. If there are no other threads of equal priority ready to run, the function returns immediately, and the thread continues execution. This behavior changed starting with Windows Server 2003.

MSDN states that the reminder of the threads time slice is relinquished to any other thread of equal priority. This is somewhat meaningless because a thread of higher priority would have been scheduled prior to the thread calling Sleep(0) and a thread with lower priority would cause the Sleep(0) to return immediately without giving anything away. Therefore Sleep(0) only has impact to threads of equal priority by default.

Purpose of Sleep(0): It triggers the scheduler to re-schedule while putting the calling thread at the end of the queue. If the queue does not have any other processes of the same priority, the call will return immediately. If there are other threads, the delay is undetermined. Note: The Windows scheduler is not a single thread, it is spread all over the OS (Details: How does a scheduler regain control when wanted?).

The detailed behavior depends on the systems timer resolution setting (How to get the current Windows system-wide timer resolution). This setting also influences the threads time slice, it varies with the system timer resolution.

The system timer resolution defines the heartbeat of the system. This causes thread quanta to have specific values. The timer resolution granularity also determines the resolution of Sleep(period). Consequently, the accuracy of sleep periods is determined by the systems heartbeat. However, a high resolution timer setting increases the power consumption.

A Sleep(period) with period > 0 triggers the scheduler and prohibits scheduling of the calling thread for at least the requested period. Consequently the calling threads time slice is interrupted. It ends immediately.

Yes, Sleep(period) with period > 0 relinquishes CPU time to other threads (if any applicable).

(Further reading: Few words about timer resolution, How to get an accurate 1ms Timer Tick under WinXP, and Limits of Windows Queue Timers).

Community
  • 1
  • 1
Arno
  • 4,994
  • 3
  • 39
  • 63