[DGD] Paranoid about MP

[Shentino]

How about a 3.0 second callout vs a 3.1 second callout?

On 8/15/06, Felix A. Croes <felix at dworkin.nl> wrote:
> Shentino <shentino at gmail.com> wrote:
>
> > I did some reading about how MP was going to be implemented and ran
> > across a concern.
> >
> > Do I need to be worried about reordering of threads?
>
> Yes.
>
>
> > More to the point:
> >
> > Barring speed differences, is it guaranteed that execution rounds in
> > MP will execute in the same order as in UP?  Or can threads commit in
> > a different order than they started in?
> >
> > Examples:
> >
> > callout 1 (0 second) is scheduled
> > callout 2 (0 second) is scheduled
> >
> > but if callout 2 is faster, it might commit before callout 1.
> >
> > Is this a concern or am I just being paranoid?
>
> The best way to make absolutely sure that callout 1 commits before
> callout 2, is to start callout 2 from callout 1.
>
> Regards,
> Dworkin
> __________________________________________
> http://mail.dworkin.nl/mailman/listinfo/dgd
>