Re: Adjustment of spinlock sleep delays

On Tue, 2003-08-05 at 18:19, Tom Lane wrote:
> Rod Taylor <rbt(at)rbt(dot)ca> writes:
> > After the first few sleeps should it add a random() element to the delay
> > time?
>
> Hmm, that's a thought --- but how big a random element?
>
> Fooling with the original idea, I'm having trouble with getting both
> plausible backoff and a reasonable number of attempts before failing.
> I tried the sequence
>
> 10 msec, 20 msec, 40, 80, ..., 1280 (1.28 sec), repeat
>
> but this only gives a couple of hundred tries before one minute has
> elapsed, which seems uncomfortably low. Maybe there's no alternative,
> though, if we want any good-sized delays in there.

How about (round to nearest 10msec):

time = oldtime + oldtime / 2 + oldtime * rand()

while (time > 1 second)
time = time - 0.80sec

This would stagger the wakeup times, and ensure a larger number of
retries -- but the times should be large enough after the first few
tries (larger than 200msec) that further backoff won't be required.