On March 7, 2002 11:49 am, William Lee Irwin III wrote:
> On Thu, Mar 07, 2002 at 09:21:19AM +0100, Andrea Arcangeli wrote:
> > then let it scale more with the ram in the machine (the
> > amount of ram used for the wait table is ridicolously small and it
> > mostly depends on the amount of the I/O, not on the amount of ram, so
> > set up a low limit instead of an high limit).
>
> The wait_table does not need to scale with the RAM on the machine
> It needs to scale with the number of waiters.
>
> 4096 threads blocked on I/O is already approaching or exceeding the
> scalability limits of other core kernel subsystems.
I think he meant that with larger ram it's easy to justify making the wait
table a little looser, to gain a tiny little bit of extra performance.
That seems perfectly reasonable to me. Oh, and there's also the observation
that machines with larger ram tend to be more heavily loaded with processes,
just because one can.
-- Daniel - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Mar 07 2002 - 21:01:02 EST